INVESTMENT MANAGEMENT

EL
Prof. Abhijeet Chandra

PT
Vinod Gupta School of Management
Indian Institute of Technology Kharagpur

N Lecture 21: Two-fund Theorem

• Markowitz Portfolio Problem

• Two-fund Theorem

EL
PT
N
• Markowitz portfolio theory
• Lagrange multiplier
• Optimal weights

EL
• Two-fund theorem

PT
N
Applications of the Markowitz Portfolio Theory
Portfolio Constraints
• In the standard portfolio theory, we want to maximize the expected return while
minimizing the standard deviation (i.e., portfolio risk), it is never ideal to hold a portfolio
on the part of the parabola below the minimum variance point.
• Hence, it is in the best interest of the investor to hold only portfolios on the upper part of

EL
the minimum-variance set (also known as efficient frontier).
E(r)

PT
Efficient Frontier
T=M

rf N
σ
Applications of the Markowitz Portfolio Theory
Portfolio Constraints
• We’re at the point in now where we’re talking about the Capital Asset Pricing Model
and the economists’ understanding of risk aversion and its consequence for the
financial sector in general.
• To understand the evolution of mutual funds as an asset class, let’s continue with the

EL
Markowitz portfolio theory and the constraints.
• Recall the n-asset Markowitz problem:
𝒏

PT
𝟏
𝒎𝒊𝒏𝒊𝒎𝒊𝒛𝒆 𝒘𝒊 𝒘𝒋 𝝈𝒊𝒋
𝟐
𝒊,𝒋=𝟏

𝒔𝒖𝒃𝒋𝒆𝒄𝒕 𝒕𝒐:
𝒏

𝒊=𝟏
N
𝒘𝒊 𝒓𝒊 = 𝒓 , 𝒂𝒏𝒅
𝒏

𝒊=𝟏
𝒘𝒊 = 𝟏 ,
Applications of the Markowitz Portfolio Theory
Portfolio Constraints 𝒏
𝟏
• Recall the n-asset Markowitz problem: 𝒎𝒊𝒏𝒊𝒎𝒊𝒛𝒆 𝒘𝒊 𝒘𝒋 𝝈𝒊𝒋
𝟐
𝒊,𝒋=𝟏
𝒏 𝒏
𝒔𝒖𝒃𝒋𝒆𝒄𝒕 𝒕𝒐: 𝒘𝒊 𝒓 𝒊 = 𝒓 , 𝒂𝒏𝒅 𝒘𝒊 = 𝟏 ,

EL
𝒊=𝟏 𝒊=𝟏

• Here, let’s allow the portfolio weight vector w to take on any real value, but

PT
sometimes we’d restrict the portfolio weights, by adding more constraints to
the above objective function.
• If we demand 𝒘𝒊 ≥ 𝟎, 𝒊 = 𝟏, 𝟐, … , 𝒏, then we do not allow short-selling.
• N
If we demand 𝒘𝒊 ∈ 𝒍𝒊 , 𝒖𝒊 , 𝒊 = 𝟏, 𝟐, … , 𝒏, then we require the portfolio
weight wi to lie between the lower and upper boundary li and ui,
respectively..
Applications of the Markowitz Portfolio Theory
Portfolio Constraints 𝒏
𝟏
• Recall the n-asset Markowitz problem: 𝒎𝒊𝒏𝒊𝒎𝒊𝒛𝒆 𝒘𝒊 𝒘𝒋 𝝈𝒊𝒋
𝟐
𝒊,𝒋=𝟏
𝒏 𝒏
𝒔𝒖𝒃𝒋𝒆𝒄𝒕 𝒕𝒐: 𝒘𝒊 𝒓 𝒊 = 𝒓 , 𝒂𝒏𝒅 𝒘𝒊 = 𝟏 ,

EL
𝒊=𝟏 𝒊=𝟏

• Here, 𝒓 is the desired level of expected rate of return (on the portfolio).

PT
𝟏
• The factor is a scaling factor in order to get nicer/simpler formula.
𝟐

N
Applications of the Markowitz Portfolio Theory
Portfolio Constraints
• To solve this optimization problem, we can use Lagrange multipliers to form the
Lagrangian L:
𝒏 𝒏 𝒏
𝟏
𝑳= 𝒘𝒊 𝒘𝒋 𝝈𝒊𝒋 − 𝝀 𝒘𝒊 𝒓𝒊 − 𝒓 − 𝝁 𝒘𝒊 − 𝟏
𝟐

EL
𝒊,𝒋=𝟏 𝒊=𝟏 𝒊=𝟏

• Here 𝝀 and 𝝁 are the Lagrange multipliers for the first and second constraints

PT
(as mentioned earlier), respectively.
• The first order conditions are given as:

𝝏𝑳
𝝏𝒘𝒊
= 𝟎, N
𝒊 = 𝟏, 𝟐, … , 𝒏,
𝝏𝑳
𝝏𝝀
=𝟎 and
𝝏𝑳
𝝏𝝁
=𝟎
Applications of the Markowitz Portfolio Theory
Portfolio Constraints
• Solving the Lagrangian, we get:
𝒏
𝝏𝑳
= 𝝈𝒊𝒋 𝒘𝒋 − 𝝀𝒓𝒊 − 𝝁 = 𝟎, 𝒊 = 𝟏, 𝟐, … , 𝒏
𝝏𝒘𝒊

EL
𝒋=𝟏
𝒏
𝝏𝑳

PT
=− 𝒘𝒊 𝒓𝒊 + 𝒓 = 𝟎
𝝏𝝀
𝒊=𝟏
𝒏
𝝏𝑳

•
𝝏𝝁
=−
N 𝒊=𝟏
𝒘𝒊 + 𝟏 = 𝟎

These equations determine the optimal vector w of portfolio weights.

Two-fund Theorem
Markowitz Portfolio Theory and its applications
• If we return to the first order conditions:
𝒏

𝝈𝒊𝒋 𝒘𝒋 − 𝝀𝒓𝒊 − 𝝁 = 𝟎, 𝒊 = 𝟏, 𝟐, … , 𝒏

EL
𝒋=𝟏 𝒏

𝒘𝒊 𝒓𝒊 − 𝒓 = 𝟎
𝒊=𝟏

PT
𝒏

𝒘𝒊 − 𝟏 = 𝟎

• N
𝒊=𝟏
Assumed that we are able to solve the problem for two different levels
of expected rate of return, 𝒓𝟏 and 𝒓𝟐 , and let the optimal weights be
indicated as 𝒘𝟏 and 𝒘𝟐 , respectively.
Two-fund Theorem
Markowitz Portfolio Theory and its applications
• Now let 𝒓 be any level of expected rate of return. Then, there should exist a unique
number α, such that:
𝜶𝒓𝟏 + 𝟏 − 𝜶 𝒓𝟐 = 𝒓

EL
namely
𝒓 − 𝒓𝟐
𝜶= 𝟏
𝒓 − 𝒓𝟐

PT
• By using the first order conditions above, we can observe that:

𝒘 = 𝜶𝒘𝟏 + 𝟏 − 𝜶 𝒘𝟐

N
is the optimal portfolio weight vector when the level of expected rate of
return is 𝒓.
Two-fund Theorem
Markowitz Portfolio Theory and its applications
• Hence, we only need to solve the Markowitz problem for two levels of expected rate
of return 𝒓𝟏 and 𝒓𝟐 . Then, we can easily get any other optimal portfolio with an
expected rate of return of 𝒓.
1. Choose 𝒓.

EL
𝒓 − 𝒓𝟐
2. Calculate the following: 𝜶= 𝟏
𝒓 − 𝒓𝟐
3. The optimal portfolio corresponding to 𝒓 is given by:

PT
𝒘 = 𝜶𝒘𝟏 + 𝟏 − 𝜶 𝒘𝟐

N
𝒓 − 𝒓𝟐 𝟏 𝒓𝟐 − 𝒓 𝟐
𝒘= 𝟏
𝒓 − 𝒓𝟐
𝒘 + 𝟏
𝒓 − 𝒓𝟐
𝒘
Two-fund Theorem
Markowitz Portfolio Theory and its applications
• The above exercise can be used to formulate the following theorem:

Any portfolio on the minimum-variance set can be written as a linear

combination of two fixed minimum-variance optimal portfolios.

EL
Any portfolio on the efficient frontier can be written as a linear
combination of two fixed efficient portfolios.

PT
N
• The Markowitz portfolio problem is basically an optimization problem with an
objective function to reduce portfolio risk, given expected rate of return.

• Using Lagrange multipliers, the objective function can be solved to obtain

EL
optimal weights that can be assigned to the individual assets in the portfolio.

• We can use the same first-order conditions to derive the two-fund

PT
theorem that suggests that any portfolio on the efficient frontier can
be written as a linear combination of two fixed efficient portfolios.

N
• Cochrane, John (2010). “Asset Pricing". New Age.
• Armerin, Fredrik (2020). “Mean-variance Analysis”., KTH.

EL
PT
N
N
PT
EL