Financial Market

Topic 2: Portfolio Theory

Bo Liu, Ph.D

Huron at Western

Jan 20, 2025

Bo Liu, Ph.D Financial Market Jan 20, 2025 1 / 68

 Mean-Variance Portfolio Theory : Part I

Returns
The return of a security is defined as the sum of the price change and any
income received over a period of time, divided by the price of the security
at the start of the time period.

Pt+1 + Xt+1
Rt+1 =
Pt
where Pt is the price and Xt is income received.

One distinguishes between gross and net return.

The net return is given by

(Pt+1 − Pt ) + Xt+1
rt+1 = Rt+1 − 1 =
Pt

We usually work with net returns expressed in %.

Bo Liu, Ph.D Financial Market Jan 20, 2025 2 / 68
 Mean-Variance Portfolio Theory : Part I

Measuring Returns
Rate of Return on Zero-Coupon Bonds:
100
r= −1
Price

100: The face value (maturity value) of the bond.

Price: The purchase price of the bond.
Effective Annual Rate (EAR):

APR n
 
1 + EAR = 1 +
n

APR: Annual percentage rate (nominal rate).

n: Compounding periods per year (e.g., n = 12 for monthly).

Bo Liu, Ph.D Financial Market Jan 20, 2025 3 / 68

 Mean-Variance Portfolio Theory : Part I

Continuous Compounding Rate

APR n
 
1 + EAR = 1+
n
As n → ∞, use the limit:

APR n
 
lim 1 + = e APR
n→∞ n

Thus:
1 + EAR = e APR
For Continuous Compounding:

APR = ln(1 + EAR)

Bo Liu, Ph.D Financial Market Jan 20, 2025 4 / 68

 Mean-Variance Portfolio Theory : Part I

Compounding
Suppose an investment yields a net return of r % each period (e.g. annual)
for n periods (e.g. years). Then, we have

Rt+n = (1 + r )n

We can calculate the effective annual return(EAR) for an n period

investment
v
u n
EAR
pn
uY
r = 1 + rt+n − 1 = t n
(1 + rt+i ) − 1
i=1

as a geometric mean of the returns per time period.

WARNING: Annual percentage rates(APR) which are simply the per
period rate r times the time intervals n in one year.

Bo Liu, Ph.D Financial Market Jan 20, 2025 5 / 68

 Mean-Variance Portfolio Theory : Part I

Example

An important concept is semi-annual returns.

To obtain these from annual returns, one needs to use

√
rsa = 1 + ra − 1.

Suppose the annual return is ra = 6%.

√
The semi-annual return is given by rsa = 1.06 − 1 = 0.02956 or 2.956%.

Bo Liu, Ph.D Financial Market Jan 20, 2025 6 / 68

 Mean-Variance Portfolio Theory : Part I

Real Versus Nominal Interest Rates

Nominal interest rate: The growth rate of your money.
Real interest rate: The growth rate of your purchasing power.

rnom = Nominal Interest Rate

rreal = Real Interest Rate
i = Inflation Rate

The relationship between real and nominal interest rates is:

rnom − i
rreal =
1+i
Note: For small values of i, the approximation holds:
rreal ≈ rnom − i
Bo Liu, Ph.D Financial Market Jan 20, 2025 7 / 68
 Mean-Variance Portfolio Theory : Part I

Interest Rates and Inflation

As the inflation rate increases, investors will demand higher nominal

rates of return.
If E (i) denotes the current expectations of inflation, then we get the
Fisher Equation:

rnom = rreal + E (i)

Bo Liu, Ph.D Financial Market Jan 20, 2025 8 / 68

 Mean-Variance Portfolio Theory : Part I

Taxes and the Real Interest Rate

Tax liabilities are based on nominal income.

rnom = Nominal Interest Rate

rreal = Real Interest Rate
i = Inflation Rate
t = Tax Rate

rnom × (1 − t) − i = (rreal + i) × (1 − t) − i = rreal (1 − t) − i × t

The after-tax real rate falls by the tax rate times the inflation rate.

Bo Liu, Ph.D Financial Market Jan 20, 2025 9 / 68

 Mean-Variance Portfolio Theory : Part I

Risk and Risk Premiums

Sources of investment risk:

Macroeconomic fluctuations
Changing fortunes of various industries
Firm-specific unexpected developments
Holding Period Return (HPR): The realized rate of return is based on
the price per share at year’s end and any cash dividends collected.

P1 − P0 + D1
HPR =
P0
HPR: Holding Period Return
P0 : Beginning price
P1 : Ending price
D1 : Dividend during period one

Bo Liu, Ph.D Financial Market Jan 20, 2025 10 / 68

 Mean-Variance Portfolio Theory : Part I

Uncertainty and Expected Returns

Looking forward the return of a security is usually uncertain. We then work
with expected return X
E [R] = Ps Rs
s
where
s are the possible states/events
Ps is the probability of return Rs
The expected return is a linear operator.
E [aR1 + bR2 ] = aE [R1 ] + bE [R2 ]
Variance (VAR):
X
σ2 = p(s) × [E (Rs ) − E (R)]2
s
Standard Deviation (STD):
√
STD = σ2
Bo Liu, Ph.D Financial Market Jan 20, 2025 11 / 68
 Mean-Variance Portfolio Theory : Part I

Scenario Returns: Example

Expected Return (E (r )):

E (r ) = (0.25 × 0.31) + (0.45 × 0.14) + (0.25 × −0.0675) + (0.05 × −0.52)

E (r ) = 0.0976 or 9.76%

Bo Liu, Ph.D Financial Market Jan 20, 2025 12 / 68

 Mean-Variance Portfolio Theory : Part I

Example VAR and STD Calculation

Example VAR calculation:

σ 2 = 0.25 × (0.31 − 0.0976)2 + 0.45 × (0.14 − 0.0976)2

+0.25 × (−0.0675 − 0.0976)2 + 0.05 × (−0.52 − 0.0976)2

σ 2 = 0.038
Example STD calculation:
√
σ= 0.038

σ = 0.1949

Bo Liu, Ph.D Financial Market Jan 20, 2025 13 / 68

 Mean-Variance Portfolio Theory : Part I

Excess Returns and Risk Premiums

Risk premium: The difference between the expected HPR and the
risk-free rate.
Provides compensation for the risk of an investment.
Risk-free rate: The rate of interest that can be earned with certainty.
Commonly taken to be the rate on short-term T-bills.
The difference between the actual rate of return and the risk-free rate
is called excess return.
Risk aversion: Dictates the degree to which investors are willing to
commit funds to stocks.

Bo Liu, Ph.D Financial Market Jan 20, 2025 14 / 68

 Mean-Variance Portfolio Theory : Part I

Learning from Historical Record

Arithmetic Average:
n n
X 1X
E (r ) = p(s)r (s) = r (s)
n
s=1 s=1
Geometric(Time-Weighted) Average:
1/n
g = TVn −1
Unbiased Estimated Variance:
n
1 X
2
σ̂ = [r (s) − r¯]2
n−1
s=1
Unbiased Estimated Standard Deviation:
v
u n
u 1 X
σ̂ = t [r (s) − r¯]2
n−1
j=1

Bo Liu, Ph.D Financial Market Jan 20, 2025 15 / 68

 Mean-Variance Portfolio Theory : Part I

Sharpe Ratio

RiskPremium
SharpeRatio =
SD(ExcessReturns)

Excess return: the difference between the actual return of a risky asset
and the actual risk-free rate.

Bo Liu, Ph.D Financial Market Jan 20, 2025 16 / 68

 Mean-Variance Portfolio Theory : Part I

Risk and Risk Aversion: Risk Aversion and Utility Values

Risk-averse investors consider only risk-free or speculative prospects

with positive risk premiums.
A portfolio is more attractive when:
Its expected return is higher, and
Its risk is lower.
Question: What happens when risk increases along with return?

Bo Liu, Ph.D Financial Market Jan 20, 2025 17 / 68

 Mean-Variance Portfolio Theory : Part I

Risk Aversion and Utility Values (continued)

We assume each investor can assign a welfare, or utility, score to
competing portfolios.
Utility Function:
1
U = E (r ) − Aσ 2
2
U: Utility value
E (r ): Expected return
A: Index of the investor’s risk aversion
σ 2 : Variance of returns
1
2 : Scaling factor

Bo Liu, Ph.D Financial Market Jan 20, 2025 18 / 68

 Mean-Variance Portfolio Theory : Part I

Investor Types

Risk-averse: Investors consider risky portfolios only if they provide

compensation for risk via a risk premium.
A > 0; A is a coefficient of risk aversion.
Risk-neutral: Investors find the level of risk irrelevant and consider
only the expected return of risk prospects.
A = 0.
Risk lovers: Investors are willing to accept lower expected returns on
prospects with higher amounts of risk.
A < 0.

Bo Liu, Ph.D Financial Market Jan 20, 2025 19 / 68

 Mean-Variance Portfolio Theory : Part I

Mean-Variance (M-V) Criterion

The selection of portfolios is based on the means and variances of

their returns.
The choice of:
The highest expected return portfolio for a given level of variance, or
The lowest variance portfolio for a given expected return.
Requirements for Portfolio A to dominate Portfolio B:
E (rA ) ≥ E (rB )
σA ≤ σB
At least one inequality is strict (to rule out indifference between the
two portfolios).

Bo Liu, Ph.D Financial Market Jan 20, 2025 20 / 68

 Mean-Variance Portfolio Theory : Part I

Trade-Off Between Risk and Return

Bo Liu, Ph.D Financial Market Jan 20, 2025 21 / 68

 Mean-Variance Portfolio Theory : Part I

Risk Aversion and Utility Values (continued)

Equally preferred portfolios will lie in the mean–standard deviation
plane on an indifference curve, which connects all portfolio points
with the same utility value

Bo Liu, Ph.D Financial Market Jan 20, 2025 22 / 68

 Mean-Variance Portfolio Theory : Part I

Return of a Portfolio

Consider a portfolio of securities where a share of Xi % is invested into

security i with uncertain return Ri .

The realized return of the portfolio is given by

N
X
RP = Xi Ri
i=1

and the expected return is given by

" N # N
X X
E [RP ] = E Xi Ri = Xi E [Ri ]
i=1 i=1

Bo Liu, Ph.D Financial Market Jan 20, 2025 23 / 68

 Mean-Variance Portfolio Theory : Part I

Variance as a Risk Measure

The variance of returns measures how much potential returns differ from
the average. It is given by
X
σ2 = Ps (Rs − E [R])2 = E [(R − E [R])2 ]
s

Also, σ is called the standard deviation.

Example: A security has either a return of 3%, 1%, or −1% with equal
probability.
The expected return is given by E (ri ) = 1%.
The variance is
1 1 1 8
σi2 = (3% − 1%)2 + (1% − 1%)2 + (−1% − 1%)2 = %2
3 3 3 3
p
or the standard deviation of returns is given by 8/3%.

Bo Liu, Ph.D Financial Market Jan 20, 2025 24 / 68

 Mean-Variance Portfolio Theory : Part I

Portfolio Example

Suppose we have two securities. The first security yields 2% for

sure(risk-free), while the second security yields 0% with probability 2/3 and
12% with probability 1/3.

The expected returns are 2% and 4% respectively.

Suppose we have a portfolio where we invest a share of 1/4 into the

risk-free asset that yields 2%, 3/4 into the risky asset.

The expected return of the portfolio is then

E [P ] = 1/4 · 2% + 3/4 · 4% = 3.5%.

Bo Liu, Ph.D Financial Market Jan 20, 2025 25 / 68

 Mean-Variance Portfolio Theory : Part I

Portfolio Example: Risky Asset and Risk-Free Asset

Let:
y : Portion allocated to the risky portfolio, P
(1 − y ): Portion to be invested in risk-free asset, F
Example
rf = 7% E (rp ) = 15%
σrf = 0% σp = 22%
Expected return on the complete portfolio:
E (rc ) = yE (rp ) + (1 − y )rf = rf + y [E (rp ) − rf ] = 7 + y (15 − 7)
Risk of the complete portfolio: σc = y · σp = 22 · y
σc 8
E (rc ) = rf + · [E (rp ) − rf ] = 7 + · σc
σp 22
E (rp ) − rf 8
Slope = =
σp 22
Bo Liu, Ph.D Financial Market Jan 20, 2025 26 / 68
 Mean-Variance Portfolio Theory : Part I

Capital Allocation Line(CAL)

Bo Liu, Ph.D Financial Market Jan 20, 2025 27 / 68

 Mean-Variance Portfolio Theory : Part I

Risk Tolerance and Asset Allocation

The investor must choose one optimal portfolio, C , from the set of feasible
choices.
Expected return of the complete portfolio:
E (rc ) = rf + y × [E (rp ) − rf ]
Variance:
σc2 = y 2 × σp2
Utility as a Function of Allocation to the Risky Asset:
1
max U = rf + y [E (rp ) − rf ] − Ay 2 σp2
y 2
FOC:
E (rp ) − rf − Ay σp2 = 0
Solve for y :
E (rp ) − rf
y∗ =
Aσp2
Bo Liu, Ph.D Financial Market Jan 20, 2025 28 / 68
 Mean-Variance Portfolio Theory : Part I

Indifference Curves different values for U and A

Bo Liu, Ph.D Financial Market Jan 20, 2025 29 / 68

 Mean-Variance Portfolio Theory : Part I

Finding the Optimal Complete Portfolio

Bo Liu, Ph.D Financial Market Jan 20, 2025 30 / 68

 Mean-Variance Portfolio Theory : Part I

Portfolio with Two Risky Assets

Suppose the fraction X1 (X2 ) is invested in security 1 (2). Then, we have

for the variance

σP2 = E [(RP − E [RP ])2 ]

= E [(X1 R1 + X2 R2 − X1 E [R1 ] − X2 E [R2 ])2 ]
= E [X12 (R1 − E [R1 ])2 + X22 (R2 − E [R2 ])2 +
+ 2X1 X2 (R1 − E [R1 ])(R2 − E [R2 ])]
= X12 σ12 + X22 σ22 + 2X1 X2 E [(R1 − E [R1 ])(R2 − E [R2 ])]
= X12 σ12 + X22 σ22 + 2X1 X2 σ12

Hence, the total variance takes into account the covariance σ12 between
the two securities.

Bo Liu, Ph.D Financial Market Jan 20, 2025 31 / 68

 Mean-Variance Portfolio Theory : Part I

Variance of a Portfolio
The derivation above is simply an application of the following formula

Var (aR1 + bR2 ) = a2 Var (R1 ) + b 2 Var (R2 ) + 2abCov (R1 , R2 )

For an arbitrary portfolio of N assets, we have

N
X N X
X N
σP2 = Xi2 σi2 + Xi Xj σij
i=1 i=1 i̸=j

where σik is the covariance between asset i and j.

We can express the relationship of the returns between two assets by their
correlation coefficient
σij
ρij = .
σi σj
Bo Liu, Ph.D Financial Market Jan 20, 2025 32 / 68
 Mean-Variance Portfolio Theory : Part I

Why does the Covariance matter?

Suppose that we have N securities with independently distributed return.

Then, ρik = 0.

Investing a fraction of 1/N we obtain

N N
X 1 X σi2 1
σP2 = (Xi2 σi2 ) = = σ̄i2
N N N
i=1 i=1

Thus, the variance of the portfolio falls as N increases unless the average
variance increases.
As long as the variance of individual assets is bounded, for N → ∞, the
variance of the portfolio goes to 0.

Bo Liu, Ph.D Financial Market Jan 20, 2025 33 / 68

 Mean-Variance Portfolio Theory : Part I

How does the correlation structure of returns influence this

result?

We have more generally

N
X N X
X N
σP2 = (Xi2 σi2 ) + Xi Xj σij
i=1 i=1 j̸=i
N N N
1 X σi2 N − 1 X X σij
= +
N N N N(N − 1)
i=1 i=1 j̸=i

or
1 2 N −1
σP2 = σ̄i + σ̄ij
N N

Bo Liu, Ph.D Financial Market Jan 20, 2025 34 / 68

 Mean-Variance Portfolio Theory : Part I

Diversification Effect

Bo Liu, Ph.D Financial Market Jan 20, 2025 35 / 68

 Mean-Variance Portfolio Theory : Part I

Insights:

Diversification reduces the variance in a portfolio. As N → ∞, the

portfolio achieves the average covariance in the market.
Individual risk of securities can be diversified away, but not the overall
market risk as given by the covariance terms.
To reduce risk in a portfolio one needs to look at the variance and
the correlation structure of the securities.

Bo Liu, Ph.D Financial Market Jan 20, 2025 36 / 68

 Mean-Variance Portfolio Theory : Part I

Alternative Risk Measures

One might be interested in risk that reflects downside risk or the risk of
extreme losses.
One measure is the semivariance which only measures downward
deviations relative to a benchmark return.
Another such measure is value at risk. The x% VaR number tells us that
we will get a return below that number in x% of the cases.
Example: The 1% VaR is -0.05 means that with probability of 1% we will
get returns below −5%.
However, this measure does not tell us anything how large these losses
could be. The expected shortfall measures the expected losses
conditional on having losses in the VaR region of returns.
normality / fat tails

Bo Liu, Ph.D Financial Market Jan 20, 2025 37 / 68

 Mean-Variance Portfolio Theory : Part II

Question

Suppose we have N securities.

Their relevant characteristics are
expected return E [Ri ]
variance σi2
correlation ρik
We can use these securities to create uncountable many portfolios with this
set of characteristics.
How can we reduce the complexity of this choice set for investors that like
higher mean returns E [RP ], but dislike higher variance σP2 ?
Idea:
Reduce the problem to “very few” portfolios to be considered.

Bo Liu, Ph.D Financial Market Jan 20, 2025 38 / 68

 Mean-Variance Portfolio Theory : Part II

Two Securities Only

We look at the case where there are only two securities D and E .
We assume that E [RD ] < E [RE ], σD < σE and a correlation coefficient
ρDE ∈ [0, 1].
Step 1: Investigate how the correlation of securities return matters for
portfolio characteristics.

Step 2: Derive the a minimum variance portfolio.

Step 3: Argue that we can restrict ourselves to the efficient frontier.

Step 4: Include the possibility of short-sales and risk-free borrowing and

lending.

Bo Liu, Ph.D Financial Market Jan 20, 2025 39 / 68

 Mean-Variance Portfolio Theory : Part II

Correlation and Portfolio Variance

Consider investing a fraction wD in D and wE = 1 − wD in E .

The portfolio’s standard deviation is a weighted average given by

σP = [wD2 σD
2
+ (1 − wD )2 σE2 + 2wD (1 − wD )σD σE ρDE ]1/2

We have

for ρDE = 1 σP = wD σD + (1 − wD )σE

q
for ρDE = 0 σP = wD2 σD2 + (1 − w )2 σ 2
D E

for ρDE = −1 σP = |wD σD − (1 − wD )σE |

We can allow for wD > 1 or wD < 0 which means that either E or D is

sold short.

Bo Liu, Ph.D Financial Market Jan 20, 2025 40 / 68

 Mean-Variance Portfolio Theory : Part II

Portfolio of Two Risky Assets

If we use equally distribute the asset: wD = wE = 0.5

Expected Return:
E (rp ) = wD E (rD ) + wE E (rE ) = 0.5 × 8% + 0.5 × 13% = 10.5%
Variance:
σp2 = wD2 σD
2
+ wE2 σE2 + 2wD wE Cov(rD , rE ) = 0.0172
Standard Deviation: √
σp = 0.0172 = 13.11%
Bo Liu, Ph.D Financial Market Jan 20, 2025 41 / 68
 Mean-Variance Portfolio Theory : Part II

Computation of Portfolio Variance from the Covariance

Matrix

Bo Liu, Ph.D Financial Market Jan 20, 2025 42 / 68

 Mean-Variance Portfolio Theory : Part II

Minimum Variance Portfolio

We can calculate a minimum variance portfolio given ρDE .

It is given by
∂σP 1 2
+ σE2 ) − σE2 + (1 − 2wD )σD σE ρDE = 0


= wD (σD
∂wD σP
The solution is given by

σE2 − σD σE ρDE σE2 − Cov (RD , RE )

wD = =
σE2 + σD2 − 2σ σ ρ
D E DE σE2 + σD2 − 2Cov (R , R )
D E

Bo Liu, Ph.D Financial Market Jan 20, 2025 43 / 68

 Mean-Variance Portfolio Theory : Part II

Mean & Variance of Portfolios: feasible (E [RP ], σP )

σP2 = wD2 σD
2
+ wE2 σE2 + 2wD wE σD σE ρDE
For ρDE = 1, the variance simplifies to:

σP2 = (wD σD + (1 − wD )σE )2

Portfolio Standard Deviation:

σP = wD σD + (1 − wD )σE

Weight wD can be derived as:

σP − σE
wD =
σD − σE

Bo Liu, Ph.D Financial Market Jan 20, 2025 44 / 68

 Mean-Variance Portfolio Theory : Part II

Mean & Variance of Portfolios: feasible (E [RP ], σP )

If we have,
σP − σE
wD =
σD − σE
The expected portfolio return is a linear function of the portfolio standard
deviation

E [RP ] = wD E [RD ] + (1 − wD )E [RE ]

   
σD E [RE ] − σE E [RD ] E [RD ] − E [RE ]
= + σP
σD − σE σD − σE

Bo Liu, Ph.D Financial Market Jan 20, 2025 45 / 68

 Mean-Variance Portfolio Theory : Part II

Mean & Variance of Portfolios

Now, let ρDE = −1.

Let wD∗ be the minimum variance portfolio.
For this case, we can derive four facts:
1 E [RP ] is a linear function of σP as in the previous case.
σE
2 For wD∗ = σE +σD we have that σP = 0.
3 For wD < wD∗ , we have σP = σE − wD (σE + σD ).
4 For wD > wD∗ , we have σP = −σE + wD (σE + σD ).

Bo Liu, Ph.D Financial Market Jan 20, 2025 46 / 68

 Mean-Variance Portfolio Theory : Part II

Mean & Variance of Portfolios

Finally, let ρDE ∈ (−1, 1).

The mean return of the portfolio E [RP ] changes linearly with wD .
The variance is a quadratic equation in wD that depends on ρDE as a
parameter.
Recall the minimum variance portfolio is given by

σE2 − σD σE ρDE σE2 − Cov (RD , RE )

wD = =
σE2 + σD2 − 2σ σ ρ
D E DE σE2 + σD2 − 2Cov (R , R )
D E

We have for the minimum variance portfolio wD∗

σE
wD∗ = 0 if and only if ρDE = σD
σE
wD∗ > 0 if and only if ρDE < σD
σE
wD∗ < 0 if and only if ρDE > σD

Bo Liu, Ph.D Financial Market Jan 20, 2025 47 / 68

 Mean-Variance Portfolio Theory : Part II

Portfolio Expected Return

Bo Liu, Ph.D Financial Market Jan 20, 2025 48 / 68

 Mean-Variance Portfolio Theory : Part II

Portfolio Standard Deviation

Bo Liu, Ph.D Financial Market Jan 20, 2025 49 / 68

 Mean-Variance Portfolio Theory : Part II

Portfolio Expected Return as a Function of Standard

Deviation

Bo Liu, Ph.D Financial Market Jan 20, 2025 50 / 68

 Mean-Variance Portfolio Theory : Part II

Insights

When short sales are not allowed, we can reduce the variance relative
to the original securities whenever ρ is NOT too much positively
correlated.
When short sales are allowed, we can always (weakly) reduce the
variance relative to the original securities.
Given a set of assets, by adjusting the weights, we can determine all
possible combinations of expected return and standard deviation for
the portfolio.
If ρ = +1 or ρ = −1, the expected portfolio return is a linear function
of the portfolio standard.
Otherwise, quadratic function.
We can extend our arguments to other, more complicated portfolios
by realizing that we can treat a portfolio just like a security.

Bo Liu, Ph.D Financial Market Jan 20, 2025 51 / 68

 Mean-Variance Portfolio Theory : Part II

The Opportunity Set of the Debt and Equity Funds and Two
Feasible CALs

Debt and equity: ρ = 0.3

Portfolio A (82% debt and 18%
equity):

E (rA ) = 8.9% σA = 11.45%

Portfolio B(70% debt and 30%

equity):

E (rB ) = 9.5% σB = 11.70%

Bo Liu, Ph.D Financial Market Jan 20, 2025 52 / 68

 Mean-Variance Portfolio Theory : Part II

Recall the sharpe ratio

Maximize the slope of the CAL for any possible portfolio, P.

The objective function is the slope:

E (rp ) − rf
Sp =
σp

Bo Liu, Ph.D Financial Market Jan 20, 2025 53 / 68

 Mean-Variance Portfolio Theory : Part II

The Sharpe Ratio: Example

Portfolio A:

E (rA ) = 8.9%, σA = 11.45%

E (rA ) − rf
SA = = 0.34
σA
Portfolio B:

E (rB ) = 9.5%, σB = 11.70%

E (rB ) − rf
SB = = 0.38
σB

Bo Liu, Ph.D Financial Market Jan 20, 2025 54 / 68

 Mean-Variance Portfolio Theory : Part II

Maximizing the Sharpe Ratio

Objective: Find the weights wD and wE that maximize the slope of the
CAL (Sharpe ratio):
E (rp ) − rf
Sp =
σp
For a portfolio with two risky assets, the expected return and standard
deviation are:
E (rp ) = wD E (rD ) + wE E (rE )
1/2
σp = wD2 σD2
+ wE2 σE2 + 2wD wE Cov(rD , rE )


Subject to the constraint wD + wE = 1, the optimization problem becomes:

E (rp ) − rf
max Sp =
wi σp

Bo Liu, Ph.D Financial Market Jan 20, 2025 55 / 68

 Mean-Variance Portfolio Theory : Part II

Maximizing the Sharpe Ratio

The weights of the optimal risky portfolio P are:

E (RD )σE2 − E (RE )Cov(RD , RE )

wD =
E (RD )σE2 + E (RE )σD
2 − [E (R ) + E (R )] Cov(R , R )
D E D E

wE = 1 − wD

Bo Liu, Ph.D Financial Market Jan 20, 2025 56 / 68

 Mean-Variance Portfolio Theory : Part II

Debt and Equity Funds with the Optimal Risky Portfolio

Optimal Risky Portfolio:

E (rp ) = 11%, σp = 14.2%

E (rp ) − rf
Sp =
σp
11% − 5%
Sp = = 0.42
14.2%

Bo Liu, Ph.D Financial Market Jan 20, 2025 57 / 68

 Mean-Variance Portfolio Theory : Part II

Determination of the Optimal Complete Portfolio

Optimal Allocation to P:

A=4
E (rp ) − rf
y=
Aσp2
11% − 5%
y= = 0.7439
4 × (14.2%)2

Bo Liu, Ph.D Financial Market Jan 20, 2025 58 / 68

 Mean-Variance Portfolio Theory : Part II

The Proportions of the Optimal Complete Portfolio

Overall Portfolio:

E (rp ) = 11%, y = 0.7439

σp = 14.2%, rf = 5%
E (roverall ) = y × E (rp ) + (1 − y ) × rf
E (roverall ) = 0.7439×11%+0.2561×5% = 9.46%
σoverall = y × σp = 0.7439 × 14.2% = 10.56%
E (roverall ) − rf 9.46% − 5%
Soverall = = = 0.42
σoverall 10.56%

Bo Liu, Ph.D Financial Market Jan 20, 2025 59 / 68

 Mean-Variance Portfolio Theory : Part II

Markowitz Portfolio Optimization Model

Security Selection:
Determine the risk-return opportunities available
Minimum-variance frontier of risky assets
All portfolios that lie on the minimum-variance frontier from the global
minimum-variance portfolio and upward provide the best risk-return
combinations
Efficient frontier of risky assets is the portion of the frontier that lies
above the global minimum-variance portfolio
Search for the CAL with the highest reward-to-variability ratio
Everyone invests in P, regardless of their degree of risk aversion
More risk-averse investors put less into P
Less risk-averse investors put more in P

Bo Liu, Ph.D Financial Market Jan 20, 2025 60 / 68

 Mean-Variance Portfolio Theory : Part II

The Efficient Frontier

Suppose we take the universe of all securities that are available for
investment.
We can derive what is called the efficient frontier.
It is guided by the following considerations.
Risk-averse investors prefer higher returns for the same variance.
Risk-averse investors prefer lower variance for the same return.
Risk-averse investors prefer diversified portfolios since they lower the
variance.
Only portfolios with returns and variance above a global minimum
variance portfolio are relevant.
This implies that the efficient frontier needs to be an increasing, concave
locus in the return-risk space.

Bo Liu, Ph.D Financial Market Jan 20, 2025 61 / 68

 Mean-Variance Portfolio Theory : Part II

The Minimum-Variance Frontier of Risky Assets

Bo Liu, Ph.D Financial Market Jan 20, 2025 62 / 68

 Mean-Variance Portfolio Theory : Part II

The Efficient Frontier of Risky Assets with the Optimal CAL

Bo Liu, Ph.D Financial Market Jan 20, 2025 63 / 68

 Mean-Variance Portfolio Theory : Part II

The Efficient Frontier of Risky Assets with the Optimal CAL

Bo Liu, Ph.D Financial Market Jan 20, 2025 64 / 68

 Mean-Variance Portfolio Theory : Part II

Capital Allocation Lines with Various Portfolios from the

Efficient Set

Bo Liu, Ph.D Financial Market Jan 20, 2025 65 / 68

 Mean-Variance Portfolio Theory : Part II

The Shape of the Efficient Frontier

Only shape (a) is possible,

while shapes (b) and (c) are
impossible due to
diversification, which implies
that portfolios must have
return-risk characteristics that
lie above a straight line.

Bo Liu, Ph.D Financial Market Jan 20, 2025 66 / 68

 Mean-Variance Portfolio Theory : Part II

Risk-free Borrowing and Lending

Investors can borrow or lend at the risk-free rate RF and invest in a risky
portfolio P to form a portfolio C .
We then have
a portfolio mean return E [RC ] = (1 − y )RF + yE [RP ]
a portfolio standard deviation σC = y σP .
Hence:  
σC σC
E [RC ] = 1 − RF + E [RP ]
σP σP
If y ∈ [0, 1], the investor lends (invests) some of its fund at the risk-free
rate.
If y > 1, the investor borrows at the risk-free rate to invest more than his
funds in the risky security.

Bo Liu, Ph.D Financial Market Jan 20, 2025 67 / 68

 Mean-Variance Portfolio Theory : Part II

Markowitz Portfolio Optimization Model

Recall:
n X
X n
σp2 = wi wj Cov(ri , rj )
i=1 j=1

Define the average variance and average covariance of the securities as:
n
2 1X 2
σ̄ = σi
n
i=1

n
1 XX
Cov = Cov(ri , rj )
n(n − 1)
i=1 j̸=i

We can then express portfolio variance as:

1 2 n−1
σp2 = σ̄ + Cov
n n
Bo Liu, Ph.D Financial Market Jan 20, 2025 68 / 68