Lesson 0B: Matrices in Finance and Python
Economics of Finance
School of Economic, UNSW1
Python
• Install the Anaconda navigator from here.
• A step-by-step instruction can be found here.
• Once the Anaconda navigator is launched, you will find
access to install Spyder, the Scientific Python Development
Environment.
• Install, and launch it.
2 / 13
Tables as Matrices
It is often desirable to think of matrices as the “inside” of
tables. For example, matrix Q might contain the values of the
following table
Bond Stock
Mon 54 21
Tue 55 18
Wed 56 27
This is a matrix with dimensions days × assets, in this case
3 × 2. In Python we’ll write
>> Q = np.array([[54,21],[55,18],[56,27]]);
3 / 13
Example: Asset allocation with investment funds
Funds Allocation, matrix A:
Fund A Fund B Fund C
Domestic Bonds 0.60 0.20 0.00
Current
Domestic Stocks 0.40 0.50 0.30
Foreign Stocks 0.00 0.30 0.70
Portfolio, vector x:
Fund A 0.20
Fund B 0.30
Fund C 0.50
4 / 13
Asset allocation
Question:
• What is the Investor’s current allocation among the three
major asset classes?
• We know that Ax = b:
0.6 0.2 0.0 0.2 0.18
0.4 0.5 0.3 0.3 = 0.38
0.0 0.3 0.7 0.5 0.44
How do we interpret b?
5 / 13
Fund Allocations
Funds Allocation, matrix A:
Fund A Fund B Fund C
Dom B 0.60 0.20 0.00
Dom S 0.40 0.50 0.30
For S 0.00 0.30 0.70
Desired Allocation, vector b:
Dom B 0.15
Dom S 0.35
For S 0.50
Question:
• What should be the portfolio in terms of fund’s investment
to obtain the desired allocation?
• Ax = b ⇒ A−1 Ax = A−1 b ⇒ x = A−1 b
6 / 13
Desirable Portfolio
Now let’s suppose there is only two funds in the market:
Fund A Fund B
Dom B 0.60 0.20
Dom S 0.40 0.50
For S 0.00 0.30
Desired Allocation, vector b:
Dom B 0.15
Dom S 0.35
For S 0.50
Question:
• What should be the portfolio in terms of fund’s investment
to obtain the desired allocation?
• not possible to answer: A is not square
7 / 13
Return and Variance
For a portfolio with each asset i weights Wi :
P∞ T
• Portfolio Return = i=1 Wi Ri =W R
P ∞ P T
• Portfolio Variance = i=1 j=1 Wi Wj σij =W σW
where
• σij is covariance of asset i with asset j;
• Ri is the return of each asset
Ri = P1 /P0 − 1
8 / 13
Example
An investor is constructing a 3-asset portfolio, denoted as X,
with allocation of $0.25 million, $0.10 million and $0.65 million
on assets A, B and C respectively. The expected returns for A,
B and C are 10%, 11% and 12% respectively.
The covariance matrix of the portfolio is given below:
Asset A B C
A 0.25 0.078 0.09
σ=
B 0.078 0.35 0.053
C 0.09 0.053 0.48
Formulate a Python command to determine the portfolio return
and standard deviation.
9 / 13
Expected return
Weight matrix and returns can be found as follows:
WT = 0.25 0.10 0.65
0.10
R = 0.11
0.12
So, the return of portfolio:
RX = WT R = 0.114 = 11.4% (1)
10 / 13
Variance
Given that:
0.25 0.078 0.09
σ = 0.078 0.35 0.053
0.09 0.053 0.48
and
0.25
W = 0.10
0.65
Variance of the portfolio can be found by:
WT σW = 0.2620
Homework: use Python to do the calculation, and then check
the next page.
11 / 13
Check point: Python
Use the following Python code:
import numpy as np
W = np.array([0.25,0.1,0.65])
R = np.array([0.10,0.11,0.12])
cov=np.array([[0.25,0.078,0.09],[0.078,0.35,0.053]
,[0.09,0.053,0.48]])
Er=np.matmul(W,R.transpose())
V=np.matmul(np.matmul(W,cov),W.transpose())
print(Er)
print(V)
12 / 13
Python Notes
• ‘import numpy as np’ to declare the usage of ‘numpy’
library
• Python uses ‘np.array([])’ to define vector
• Matrix is writen as nesting of vectors, i.e.,
‘np.array([[],[],...,[]])’
• Matrix multiplication: np.matmul()
• Python does not allow multiplication of a sequence of more
than two matrices: np.matmul(A,B,C) is invalid
• Nesting, however, is more than welcome:
np.matmul(np.matmul(A,B),C).
13 / 13
