Investments

Lecture 3.2 (Markowitz)

MSc. Finance
Dr. Anna Bayona
Table of Contents 2

1 –The relationship between portfolio risk and return

2 –Markowitz model of portfolio choice
3 –The practice of investments
 3

1
The relationship between
portfolio risk and return
2 4

Portfolio return and

risk
Example of expected
return for a portfolio
with 2 assets
• Suppose that we have two assets in our portfolio, 1 and 2,
with weights w1 and w2 , and expected returns E(R1) and E(R2), respectively.

• Then, we note that 𝒘𝟏 + 𝒘𝟐 = 𝟏.

• The expected return on the portfolio is

𝐸 𝑅! = 𝑤" 𝐸 𝑅" + 𝑤# 𝐸 𝑅#
2 5

Portfolio return and

risk
Expected portfolio
return with N assets
• Suppose that the assets in a portfolio, including cash, account for 100%, that is,
the sum of the weights for each individual asset in the portfolio must add up to
&
1. That is, for individual assets 𝑖 = 1, … , 𝑁, then ∑#$% 𝑤# = 1 , where 𝑤# is the weight
of asset i portfolio P.
• When several individual assets are combined into a portfolio, we can compute
the expected portfolio return as a weighted average of the expected returns
in the portfolio. That is, for individual assets 𝑖 = 1, … , 𝑁, the expected return of
the portfolio, 𝐸 𝑅' , is

&
∑$%" 𝑤$ 𝐸(𝑅$ ),
𝐸 𝑅! =

where 𝐸(𝑅! ) is the expected returns of asset i.

2 6

Portfolio risk for 2

Portfolio return and
risk

assets
# # #
• For two assets with variance 𝜎" and 𝜎# , the portfolio variance, 𝜎$ , becomes:

# # # # #
𝜎! = 𝑤" 𝜎" + 𝑤# 𝜎# + 2𝑤" 𝑤# 𝐶𝑜𝑣(𝑅" , 𝑅# ),

where 𝐶𝑜𝑣(𝑅" , 𝑅# ) is the covariance of returns 𝑅" and 𝑅# , and we can re-write it as:

# # # # #
𝜎! = 𝑤" 𝜎" + 𝑤# 𝜎# + 2𝑤" 𝑤# 𝜌"# 𝜎" 𝜎# ,

where 𝜌"# is the correlation between the two returns, and 𝜎! the standard deviation of
asset i.
• The portfolio standard deviation is given by the square root of the portfolio’s varianc,
and it is the usual measure of risk.
2 7

Portfolio return and

risk Covariance and
Correlation
• The covariance in the formula for the portfolio variance can be expanded as

𝐶𝑜𝑣 𝑅" , 𝑅# = 𝜌"# 𝜎" 𝜎#

• The correlation coefficient 𝜌"# represents the consistency or tendency for two
investments to act in a similar way. The correlation coefficient ranges from -1
to +1. Let us consider the various cases:
• 𝝆𝟏𝟐 = +𝟏. Returns of the two assets are perfectly positively correlated. Assets 1 and 2
move together 100% of the time. 𝜎#$ = 𝑤%$ 𝜎%$ + 𝑤$$ 𝜎$$ + 2𝑤% 𝑤$ 𝜎% 𝜎$ = (𝑤% 𝜎% + 𝑤$ 𝜎$ )$ ,
• 𝝆𝟏𝟐 = −𝟏. Returns of the two assets are perfectly negatively correlated. Assets 1 and 2
move in opposite directions 100% of the time.
• 𝝆𝟏𝟐 = 𝟎. Returns of the two assets are uncorrelated. The movement of asset 1 provides
no prediction regarding the movement of asset 2.
# #
• Notice that for 𝜌"# < +1 then 𝜎$ < (𝑤" 𝜎" + 𝑤# 𝜎# ) .
2 8

Portfolio return and

risk
Illustration of the
correlation coefficient
2 9

Portfolio return and

risk
Example

• Suppose that we have 2 assets with the following characteristics:

• Asset 1. Return 10% and standard deviation 20%
• Asset 2. Return 5% and standard deviation 10%
• Correlation between assets 1 and 2: 0
• Let us form 3 portfolios with these 2 assets and calculate the portfolio returns
and standard deviations.
PORTFOLIO WEIGHT IN WEIGHT IN PORTFOLIO PORTFOLIO
ASSET 1, % ASSET 2, % RETURN (%) STANDARD
DEVIATION (%)
X 25 75 6.25 9.01
Y 50 50 7.50 11.18
Z 75 25 8.75 15.21
2 10

Portfolio variance
Portfolio return and
risk

for N risky assets

• We can now write the portfolio variance for N risky assets as:

# & # #
∑$%" 𝑤$ 𝜎$ &
∑$,(,",$)( 𝑤$ 𝑤( 𝐶𝑜𝑣(𝑅$ , 𝑅( ),
𝜎! = +

• For N risky assets, matrix notation can be used to simplify the computations. This
can be also implemented in computer programs to facilitate the calculations.
2 11

Portfolio return and

risk
Sharpe ratios around the world:
1900 to 2020
Sharpe ratio: Reward –
to-volatility ratio.

Source: Bodie, Kane

and Marcus (2023)
3 12

Portfolio risk and

return
Example of a
portfolio’s
opportunity
set for 2 risky
assets

Source: Bodie, Kane and Marcus (2014)

3 13

Portfolio risk and

return The power of
diversification
• Risk averse investors are interested in reducing risk preferably without reducing
return. In other cases, investors may accept a lower return if it will reduce the
chance of catastrophic losses.
• We have seen the importance of correlation and covariance in managing risk.
• Correlation is key in the diversification of risk if correlation is less than +1.
• Approaches to diversification:
• Diversify with asset classes
• Diversify with index funds
• Diversify among countries
• Diversify by not owning too much of your employer’s stock
• Buy insurance for a risky portfolio
3 14

Portfolio risk and

return
Investment opportunity
set of risky assets and
the efficient frontier
All points on the curve and
points to the right of the
curve are attainable by a
combination of one or more
investible assets.

Source: Bodie, Kane and Marcus (2014)

3 15

Portfolio risk and

return
Interpretation

• When we consider a higher number of investable assets, the number of combinations

increases. When we consider all investable assets, then we can construct an opportunity
set of investments, which consists of all available investable sets.
• From now on, we assume that all investable assets are included in the investment opportunity set.

• The minimum-variance frontier represents the lowest possible variance that is attainable
given an expected return.

• The global minimum-variance portfolio is the portfolio of risky assets with the lowest
variance on the minimum-variance frontier.

• The efficient frontier of risky assets is the upper part of the hyperbola from the global
minimum-variance portfolio and upward. It provide the best risk-return combinations.
3 16

Portfolio risk and

return Combinations of a
riskless asset and a
risky asset
• Consider a portfolio P’ of 2 assets: a risky one and a risk free one.
• The risk-free asset has return 𝑅& and zero risk. Usually this is a debt security with no default
risk, no inflation risk, no liquidity risk, no interest risk and no risk of any other kind.
• The risky asset i has expected return 𝐸 𝑅' and 𝜎'$ , such that 𝐸 𝑅' > 𝑅& .
• Let us start draw the investment opportunity set. The formulae that we need are:
𝐸 𝑅$% = 𝑤& 𝑅& + 𝑤! 𝐸 𝑅!
# # #
𝜎$%= 𝑤! 𝜎! ,
such that 𝑤! = 1 − 𝑤&
• This is represented by the capital allocation line (CAL), which represents the portfolios
available to an investor when combining a risk-free asset and risky assets.
3 17

Portfolio risk and

return
Capital allocation line
• We can derive its equation from the previous
equations and noting that from the variance
(!"
equation we can write: 𝑤' = , and
(#
substituting it into the expected return
equation we derive:
(+ ,# -,$ )
𝐸 𝑅#) = 𝑅& + 𝜎#)
(#

• In other words, the CAL is a straight line with

(+ ,# -,$ )
intercept 𝑅& and slope , which is the
(#
additional return required for every
increment of risk (market price of risk or
Sharpe ratio).
• To the NE of the bold point, we can achieve it
by borrowing at the risk-free rate and
investing it in the risky asset.
1 18

2
Markowitz Model of
Portfolio Choice
2 19

Markowitz
Motivation

• The goal is to choose the optimal portfolio weights that provide the optimal
trade-off between expected return and variance (risk) of the portfolio for an
investor with a given risk tolerance.

• For calculating the optimal portfolio according, we need:

1. Expected returns, variances and covariances of assets.

2. Investor’s risk preferences
2 20

Markowtiz
Assumptions and
objective
• The model is based on several assumptions:

• A single period horizon (static model).

• Investors are risk averse and maximize expected utility by using the mean-
variance criterion.
• No transaction costs or taxes.
• Assets are infinitely divisible.
• Investors can invest in all assets without restrictions.
• Investors cannot influence prices (prices are given; competitive prices).

• Given a target mean expected return of the portfolio, Markowitz (1952)

characterizes the efficient portfolio by choosing the weights to invest in
each asset that minimise the variance of the portfolio.
2 21

Markowitz
The two-fund
separation property
• Portfolio choice problem may be separated into two independent tasks (Tobin-
Markowitz)
• Investment decision: Determination of the optimal risky portfolio is purely
technical (regardless of an investor’s risk aversion).

• Financing decision. The allocation of the complete portfolio to risk-free

asset versus the risky portfolio depends on personal preference with
respect to risk-reward. Each investor’s portfolio on the CAL is determined.
Portfolios beyond the optimal risky portfolio are obtained by borrowing at
the risk-free rate.
•More risk averse investors give more weight on the risk-free asset
•Less risk averse investors give more weight on the risky portfolio.
2 22

Markowitz
Risk averse, neutral
and seeking

• It is possible to create a complete portfolio by splitting investment funds between risk-

free and risky assets

• Let
•y = Portion allocated to the risky portfolio, P
•(1 - y) = Portion to be invested in risk-free asset, F F
P
2 23

Markowitz
The optimal risky
portfolio with a riskless
asset and many risky assets
• All the portfolios on the efficient frontier are
candidates for being combined with the risk-free
asset.
• We can try to move further NW but these
portfolios are not attainable. We can move SE
but these are dominated by P.
• Portfolio P is the optimal risky portfolio at
which the CAL is tangent to the efficient
frontier.
• It has the highest attainable Sharpe ratio (slope
' (! )("
of the CAL which is ).
*!
3 24

Application
Which portfolio do we
choose given a level of risk
aversion?
• Suppose we consider a portfolio of 2 assets: a risky asset and a riskless asset.

• The risk-return combinations available to an investor are described by the

capital allocation line (CAL).

• Which of these portfolios should be chosen by an investor?

• The indifference curves give us the risk-return preferences of an individual.
• The capital allocation line (CAL) gives us the set of feasible investments.
• The tangency point between the two lines gives us the optimal portfolio for an investor.
3 25

Application

Portfolio selection
for an investor
• Four indifference curves are shown
for the same investor and the CAL.
• The second indifference curve
starting from the top is tangential to
the CAL.
• The curve on the NW is not
achievable with the available assets.
• The curves on the SE are dominated
by the second curve.
• Thus the optimal portfolio C is
selected. It is the tangency point
between the CAL and the indifference
curve.
Source: Bodie, Kane and Marcus (2014)
3 26

Application

Portfolio selection
for two investors
with different risk preferences
• Two investors A and B whose
preferences are described by curves IA
and IB. Investor B has a higher risk
tolerance than investor A.
• The two optimal portfolios for each
investor are different: B has a
higher proportion of the risky asset
compared to A.
3 27

Portfolio risk and

return Complete portfolio
characteristics

• Consider the complete portfolio C, which consists of:

• The risk-free asset has return 𝑅& and zero risk. We invest a proportion 1 − 𝑦 of our wealth.
• The optimal risky risky asset P has expected return 𝐸 𝑅# and variance 𝜎#$ . We invest a
proportion 𝑦 of our wealth.

• The expected return and risk of the complete portfolio are:

𝐸 𝑅+ = (1 − 𝑦)𝑅& + 𝑦𝐸 𝑅$
𝜎+ = 𝑦𝜎$
2 28

Markowitz
Determining the
optimal complete
portfolio C

Source: Bodie, Kane

and Marcus (2014)
2 29

Markowitz
Finding the optimal
complete portfolio
with n>2 risky assets
• With more than two assets (n>2), the algebra of the efficient frontier becomes
cumbersome since it depends on many covariance terms.

• We can assume that investors choose portfolios to minimize the portfolio’s risk
subject to a target rate of return.

• These portfolios lie on the boundary of the investment opportunity set and above
the global minimum variance portfolio.

• Matrix notation can simplify many of the computations. This can be easily translated
into computer programs such as Excel, R, Matlab, etc.
2 30

Markowitz
Limitations
• The problem with Markowitz’s approach is that the number of estimates required is
very large. In a portfolio with N assets, we need to estimate N(N-1)/2 correlations!

• Sometimes it produces extreme portfolios with very large weights on some assets and
large negative weights on others. One way to solve this is to add portfolio constraints
– which can be easily added.

• The portfolio weights are very sensitive to small changes in expected rates of return.

• Errors in estimated the variance-covariance matrix can also have a large impact on
the optimal portfolio.

• The theory is based on expected utility. Behavioral extensions.

 31

3
The practice of investments
 32

1
The investment process
3

The Investment Management Process is Dynamic

The Investment Process

PLANNING

EXECUTION

FEEDBACK
Source: NYU, CFA Institute
4

The Investment Management Process

Planning: Types of Investors & Return Requirements & Risk

Source: Bodie, Kane and Marcus (2014)

5

The Investment Management Process

Determination of Portfolio Policies

Source: Bodie, Kane and Marcus (2014)

ESG considerations are becoming more important in portfolio planning
and construction. Examples: exclusionary screening, best-in class
selection, active ownership, thematic and impact investment, and ESG
integrated in security analysis.
6

The Investment Management Process

Top down and bottom up approaches

Source: CFA
7

The Investment Management Process

Portfolio construction

Long-term risk and

return goals and
stable

Shorter-term and
market timing

Source: JP Morgan
1 38

Expected return
and risk of a single
asset
Asset class returns

Source: Visual Capitalist

1 39

Expected return
and risk of a single
asset
Forecast returns for major asset classes

Source: The visual

capitalist
1 40

Expected return
and risk of a single
asset
Asset class correlation matrix

Source: Visual Capitalist

1 41

Expected return
and risk of a single
asset
Harvard Management Company
policy portfolio

Source: Harvard
Management Company
(2010) case
Investments
Lecture 3.2
MSc. Finance
Dr. Anna Bayona