First Semester, 2024-2025

THE UNIVERSITY OF HONG KONG

DEPARTMENT OF STATISTICS AND ACTUARIAL SCIENCE

STAT3904 CORPORATE FINANCE FOR ACTUARIAL SCIENCE

Tutorial 4: Mean-Variance Portfolio Theory (Part I)

1 Key Learning Points

In the test and examination, candidates are expected to:

LP1 Understand the definition and importance of beta.

LP2 Find the global minimum variance portfolio.
LP3 Find another portfolio on the minimum variance set.
LP4 Derive the equation of the efficient frontier

(a) when there are two risky assets by mixing;

(b) when there are n ≥ 3 risky assets by mixing the minimum variance portfolio and
another efficient portfolio. (will be discussed later)

2 Review of Key Concepts

2.1 Market risk and Beta

• While diversification can eliminate specific risks in a portfolio, the market risk still re-
mains. Base on the assumption that every investor can construct a well-diversified port-
folio, then the risk of the portfolio is determined by the market risk, measured by the
concept of beta.
• Definition: The beta of an individual security measures its sensitivity to market move-
ments, defined as the covariance of the individual security and the market portfolio divided
by the variance of the market portfolio. Notationally,
σim
βi = 2 .
σm
• The importance of beta: For a well-diversified portfolio, only the market risk remains,
and the portfolio beta (usually the weighted average of the beta of individual securities)
will describe the risk well (actually it is just a ratio of market risk depending on the value
of portfolio beta).
• We will come back on Beta later after the capital asset pricing model (CAPM).
S&AS: STAT3904 Corporate Finance for Actuarial Science 2

2.2 Two-asset Market

• The two-asset case already captures the essence of mean-variance portfolio theory. Com-
plicated multi-asset problems can sometimes be solved by reducing them to the two-asset
case.

• Fundamental Equations: The whole analysis begins with the following equations:

(Mean) µ P = x1 µ 1 + x2 µ 2
(Variance) σP2 = x21 σ12 + 2x1 x2 ρσ1 σ2 + x22 σ22
(Weight Constraint) x1 + x 2 = 1
SOA
Exercise 1. (Warm-up Exercise) You are given the following with respect to a portfolio Course
consisting of two mutual funds: 6 Spring
2003
Mutual Fund Weight Variance
Stock 25% 100
Bond 75% 36

The correlation coefficient between the fund returns is −0.5. Calculate the variance of
the portfolio.
Solution. Denote Stock and Bond as assets A and B respectively. The portfolio variance
can be calculated as follows:

σP2 = x2A σA2 + x2B σB2 + 2xA xB ρσA σB

1
= 0.252 × 100 + 0.752 × 36 + 2 × 0.25 × 0.75 × (−0.5) × (100 × 36) 2
= 15.25.

How to Find the Minimum Variance Set (MVS) & Efficient Frontier (EF)? The
nicest thing about the two-asset market is that the minimum variance set can simply be
derived by mixing the two given assets. One can simply solve the above set of equations
and express x1 , x2 in terms of µ’s and σ’s, eventually leading to the relation between
µP and σP . The detailed steps are as follows:
1. From the weight constraint equation, x2 = 1 − x1 .
µP − µ2
2. Solving the first equation for x1 , we get x1 = .
µ1 − µ2
−µ2
3. Substituting x1 = µµP1 −µ2
and x2 = 1 − x1 into the second equation, we obtain the
relationship between σP and µP :
 2  2
2 µP − µ2 2 µP − µ2 (µP − µ2 )(µ1 − µP )
σP = σ1 + 1 − σ22 + 2 ρσ1 σ2 .
µ1 − µ2 µ1 − µ2 (µ1 − µ2 )2

• Minimum Variance Portfolio (MVP):

– Finding the minimum variance portfolio means determining the portfolio composi-
tion x = (x1 , x2 )T (NOT µP and σP !) that minimizes σP2 .
S&AS: STAT3904 Corporate Finance for Actuarial Science 3

– By a simple calculus argument on the variance formula together with the weight
constraint (see page 1, Chapter 8), we obtain1

σ22 − ρσ1 σ2 σ12 − ρσ1 σ2

x∗1 = , x∗2 = 1 − x∗1 = . (1)
σ12 + σ22 − 2ρσ1 σ2 σ12 + σ22 − 2ρσ1 σ2

This is a handy expression that is worthwhile to remember ; even in a multi-asset

market, Equation (1) allows us to relate the global minimum variance portfolio to
two portfolios lying on the minimum variance set.

• When a risk-free asset is introduced to the market, rendering the possibility of borrowing
and lending, the efficient frontier can be further improved. By the one-fund theorem,
the new frontier is graphically the tangent line of the original one, and the tangent point
corresponds to the tangency portfolio.

2.3 n-asset Market (To be continued)

• Matrix notation will become handy in the multi-asset case.

– Portfolio rate of return: RP = xT R, where R = (R1 , . . . , Rn )T is the vector of

returns of the n assets
– Expected rate of return: µP = E[xT R] = xT E[R] = xT µ
– Variance of rate of return: σP2 = Var(xT R) = Cov(xT R, xT R) = xT Cov(R, R)(xT )T =
xT Σx

3 Problems

Attempt ALL THREE questions. Marks for past paper questions are shown in square brack-
ets.

1. Explain in words two different methods to derive the equation of the minimum variance set.

Solution. (a) Fix a number µ0 ∈ R. Determine the portfolio with the smallest variance
subject to the constraint that its mean equals µ0 and calculate its standard deviation.
Repeat the same analysis for different values of µ0 .
(b) In the second method, it suffices to find the global minimum variance portfolio (by
normalizing Σ−1 1) and another minimum variance portfolio (by normalizing Σ−1 µ). By
the two-fund theorem, mixing these two portfolios will capture all minimum variance
portfolios.

1
Observe that x∗ = (x∗1 , x∗2 )T is independent of µ = (µ1 , µ2 )T . Hence any change in µ will not affect x∗ .
S&AS: STAT3904 Corporate Finance for Actuarial Science 4

2. Mean-Variance Portfolio Analysis STAT2807

Assume that there are only two risky stocks. Portfolio A has an expected return rate (µA ) 09-10
of 20% and standard deviation (σA ) of 40%. Portfolio B has an expected return rate (µB ) of Test
50% and standard deviation (σB ) of 100%. The correlation coefficient between the returns
of portfolio A and B is 0.8.
(a) Find the global minimum variance portfolio. Calculate the mean and variance of its
return.
(b) Suppose the market has no risk-free assets and it is known that both portfolios are on
the efficient frontier. Derive the equation of the efficient frontier. [8 marks]
(c) If we add a risk free asset in the market and assume that the risk-free rate is 5%, find
the efficient portfolio of risky assets, and derive the equation of the efficient frontier.
[7 marks]

[Total: 15 marks]

Solutions. (a) The global minimum variance portfolio consists of

σB2 − ρσA σB 1 − 0.8(0.4)(1) 17
xA = 2 2
= 2
=
σA + σB − 2ρσA σB 1 + 0.4 − 2(0.8)(0.4)(1) 13
4
units in Asset A and 1 − xA = − 13 unit in Asset B. The expected rate of return is thus
 
17 4 7
µ P = xA µ A + xB µ B = × 0.2 + − × 0.5 = ,
13 13 65
and its variance is
36
σP2 = x2A σA2 + x2B σB2 + 2xA xB ρσA σB =
.
325
(b) Let α be the weight of Asset A in the portfolio, then we have
µ = 0.2α + 0.5(1 − α),
σ 2 = 0.16α2 + (1 − α)2 + 2 × 0.4 × 1 × 0.8α(1 − α).
Eliminating α yields
 2  2   
2 0.5 − µ µ − 0.2 0.5 − µ µ − 0.2
σ = 0.16 + + (0.64) ,
0.3 0.3 0.3 0.3
7
with µ ⩾ 65 .
(c) Let α∗ be the weight of Asset A in the tangency portfolio. Using the result in Page
10 of Chapter 8 lecture notes, the efficient portfolio of risky assets can be solved by
normalizing
      
−1 µ1 − rf 1 −0.32 0.15 0.006
Σ ∝ = ,
µ2 − rf −0.32 0.16 0.45 0.024
which gives α∗ = 0.2 and 1 − α∗ = 0.8. Routine calculation then yields µM = 0.44 and
2
σM = 0.7488 for the tangency portfolio. Hence the equation of the efficient frontier,
which is the capital market line, is
µM − rf
µ − rf = σ ⇒ µ = 0.05 + 0.4507σ, σ ≥ 0 .
σM
S&AS: STAT3904 Corporate Finance for Actuarial Science 5

3. Some Facts of the Covariance of Two Portfolios

(a) Suppose that there are n risky stocks in the market with returns R1 , . . . , Rn respectively. STAT2807
You are given: 09-10

• Portfolio 1 puts a weight of xi in stockP i, so that we can represent portfolio 1 by a Test

column vector x = (x1 , . . . , xn )T with ni=1 xi = 1.
• Similarly, Portfolio 2 puts a weight ofP yi in stock i, so that it is represented by a
column vector y = (y1 , . . . , yn ) with ni=1 yi = 1.
T

• Σ = (σij ) is the n × n variance-covariance matrix of the returns.

Express the covariance between the return of Portfolio 1 and the return of Portfolio 2
using matrix notation. [5 marks]

Solution. The covariance between the returns of Portfolios 1 and 2 is given by

n n
! n Xn
X X X
Cov xi Ri , yj Rj = xi yj Cov(Ri , Rj )
i=1 j=1 i=1 j=1
n X
X n
= xi yj σij
i=1 j=1

= xT Σy.

The final result is called the bilinear form associated with Σ, and its special case when
x = y recovers the variance of the portfolio return: Cov(xT R, xT R) = Var(xT R) =
σP2 = xT Σx.

(b) Prove that the covariance between the returns of any portfolio and the global minimum-
variance portfolio is the same. Identify that common covariance.

Proof. Let P and Q be the global minimum variance portfolio and an arbitrary portfolio
respectively and denote their rates of return by RP and RQ . To determine a minimum
variance portfolio in a market consisting only of Portfolios P and Q, the weight of Q
must be 0 due to its arbitrariness. Therefore, the formula in Section 2.2 implies

σP2 − Cov(RP , RQ )
0= ,
σP2 + σQ2
− 2Cov(RP , RQ )

or Cov(RP , RQ ) = σP2 for any Q.

Remark 1. Note that this result holds regardless of whether the arbitrary portfolio Q
is efficient or not.
Remark 2. As a by-product, what is the correlation coefficient between RP and RQ ,
ρP Q ?

********** END OF TUTORIAL 4 **********