ECN 2322

MATHEMATICS FOR ECONOMICS II

LECTURE 1
EIGENVALUES AND EIGENVECTORS

Mathematics for Economics II Department of Economics

 Eigenvalues and vectors
• Eigenvalues are sometimes referred to as latent roots or
characteristic roots.
• The word eigen is derived from the German word which
means own. Hence eigenvalues are simply a given matrix’s
own values and the same goes for eigen vectors.
• Let 𝐴 be a matrix of dimensions 𝑛 × 𝑛 and 𝑥 be a non-null
column vector with elements 𝑛. Then
𝐴𝑥 = 𝜆𝑥
• Then 𝜆 is called the eigenvalue.

Mathematics for Economics II Department of Economics 2

• 𝐴𝑥 = 𝜆𝑥 can also be expressed as
𝐴 − 𝐼𝜆 𝑥 = 0
• For the solution to this equation to exist 𝐴 − 𝐼𝜆 must be
singular. That is,

𝑎11 − 𝜆 𝑎12 − 𝜆 ⋯ 𝑎1𝑛 −

𝑎21 − 𝜆 𝑎22 − 𝜆 ⋯ 𝑎2𝑛 − 𝜆
𝐴 − 𝐼𝜆 = =0
⋮ ⋮ ⋮ ⋮
𝑎𝑛1 − 𝜆 𝑎𝑛1 − 𝜆 ⋯ 𝑎𝑛𝑛 − 𝜆

Mathematics for Economics II Department of Economics 3

• When we expand this determinant the resultant polynomial which is
called the characteristic polynomial will be in terms of 𝜆.
• The roots of the characteristic polynomial are known as eigenvalues.
• A vector 𝑥 associated with each eigenvalue is known as eigen vector.
• A unique solution will not be found due to the fact that the
coefficient matrix is singular. The equations will be multiples of each
other and as such there be linear dependence.
• The vector obtained will only have direction but no magnitude.
• In order to ensure that our vector has magnitude we normalise by
using a magnitude of one.

Mathematics for Economics II Department of Economics 4

• Thus the normalised condition used to obtain a unique
vector is
𝑥12 + 𝑥22 = 1
• Note that the 𝑥 ′ 𝑠 are not vectors but elements of the vector
while 𝑣 will denote the vector.

In-class exercise 1.1

4 2
Given a square matrix 𝐴 = , find the eigenvalues and the
2 7
respective eigen vectors.
Mathematics for Economics II Department of Economics 5
 Diagonalization of a matrix
• Recall in mathematics for economics I we defined a diagonal
matrix as a type of square matrix in which all nondiagonal
elements are equal to zero as shown below:
𝑎11 0 0
𝑨 = 0 𝑎22 0
0 0 𝑎33
• Diagonal matrices play a critical role in simplifying
mathematical operations of economics. They easy both to
multiply and add and their usefulness will seen later on.
Mathematics for Economics II Department of Economics 6
• We can use eigenvalues and eigen vectors to diagonalize a
matrix.
• Diagonalization is the process of transforming an ordinary
matrix into a diagonal matrix or form.
• For distinct eigenvalues, the matrix 𝐴 can be transformed
into a diagonal matrix by using a transformation matrix 𝑇.
That is
𝑇 = 𝑣1 𝑣2 ⋯ 𝑣𝑛
• For an 𝑛 × 𝑛 matrix 𝐴 its diagonal matrix 𝐷 is given by:

Mathematics for Economics II Department of Economics 7

 𝐷 = 𝑇 −1 𝐴𝑇
• Where
𝜆1 0 0 0
0 𝜆2 0 0
𝐷=
0 0 ⋱ 0
0 0 0 𝜆𝑛
• Where 𝜆′ 𝑠 are the respective eigenvalues associated with
the eigen vectors in the transformation matrix 𝑇.

Mathematics for Economics II Department of Economics 8

 In-class exercise 1.2
4 1
Given a matrix 𝐴 = , find the eigenvalues, eigen vectors
3 2
and the corresponding diagonal matrix 𝐷.

Mathematics for Economics II Department of Economics 9

 Non-negative square matrices
1. Positive matrix is one where all of its elements are positive.
A positive matrix does not include zero.
2. Non-negative matrix on the other hand is a matrix whose
elements are zeroes and any other positive numbers. That
is, it is a matrix that has a zero in it but all other elements
are positive.
3. Semi-positive matrix is a non-negative matrix with each
row and column having alteast one positive element.

Mathematics for Economics II Department of Economics 10

 Symmetric and non-symmetric matrices
1. Upper-triangular matrix. In this case entries where number
corresponding to the row is greater than the number
associated with the column on an element are equal to zero.
That is, 𝑎𝑖𝑗 = 0 whenever 𝑖 > 𝑗. This matrix is usually a
square.
2. Lower-triangular matrix. In this case we have a zero entry
where a number corresponding to a row is less than the
number associated with the column on an element are
equal to zero. That is, 𝑎𝑖𝑗 = 0 whenever 𝑖 < 𝑗. Usually a
square matrix.

Mathematics for Economics II Department of Economics 11

3. Symmetrical Matrix. The symmetrical matrix relate to the
mirror effect along the main diagonal. This is where 𝐴𝑇 =
𝐴. That is, 𝑎𝑖𝑗 = 𝑎𝑗𝑖 , for all 𝑖 𝑎𝑛𝑑 𝑗. An example of a
symmetrical matrix is the identity matrix.
4. Skew symmetrical. This is where there is a mirror effect
along the main diagonal but the elements even though the
same they change signs, from positive to negative. That is,
𝑎𝑖𝑗 = −𝑎𝑗𝑖 , for all 𝑖 𝑎𝑛𝑑 𝑗. The matrix below is an example:
0 1 −3
𝐴 = −1 0 2
3 −2 0

Mathematics for Economics II Department of Economics 12

The Carley Hamilton Theorem
• For any square matrix 𝐴, the characteristic equation is a
polynomial in lamda (𝜆). The equation is denoted by 𝑓 𝜆 =
0.
• Carley-Hamilton theorem says that any such matrix satisfies
its own characteristic equation. If the matrix 𝐴 is substituted
in place of 𝜆 in the equation, the equation still holds.
2 4
• For instance matrix 𝐴 = its characteristic
−2 8
polynomial is given by:
Mathematics for Economics II Department of Economics 13
 𝑓 𝜆 = 𝐴 − 𝜆𝐼 = 0

2−𝜆 4
=0
−2 8−𝜆
= 2−𝜆 8−𝜆 +2×4=0
= 𝜆2 − 10𝜆 + 24 = 0

2 4 2 4 2 4 1 0 0 0
𝑓 𝐴 = − 10 + 24 =
−2 8 −2 8 −2 8 0 1 0 0

Mathematics for Economics II Department of Economics 14

Decomposable and indecomposable matrices
• A matrix 𝐴 with dimensions 𝑛 × 𝑛 is said to be
decomposable if it is possible to interchange or rearrange its
rows and columns in such a way as to obtain a matrix of the
form
መ 𝐴11 𝐴12
𝐴= , 𝑤ℎ𝑒𝑟𝑒
𝐴21 𝐴22
𝐴11 = 𝑘 × 𝑘 𝑚𝑎𝑡𝑟𝑖𝑥, 𝐴22 = (𝑛 − 𝑘) × (𝑛 − 𝑘) 𝑚𝑎𝑡𝑟𝑖𝑥
𝐴21 = (𝑛 − 𝑘) × 𝑘 𝑚𝑎𝑡𝑟𝑖𝑥, 𝐴12 = 𝑘 × (𝑛 − 𝑘) 𝑚𝑎𝑡𝑟𝑖𝑥
Mathematics for Economics II Department of Economics 15
• From above we can see that decomposing a matrix is actually
breaking the matrix into four submatrices.
• The requirement for decomposing a matrix is that matrix 𝐴21
should be a null matrix after decomposition.
• Therefore, a matrix is decomposable if after the
rearrangement to form submatrices, the 𝐴21 will be a null
matrix (𝐴21 matrix with zeroes) otherwise it is
indecomposable.
• Matrices are decomposable if they have a zero in every row
or column.
Mathematics for Economics II Department of Economics 16
• In economics, a group of industries or sectors are said to be
indecomposable if every industry in the economy is linked,
directly or indirectly through others, to all other sectors.
In-class exercise 1.3
Decompose the following matrix
1 0 2
𝐴= 3 4 5
6 0 7

Mathematics for Economics II Department of Economics 17

 Stochastic matrices
• Assume we have two states 𝑖 and 𝑗. In agriculture state 𝑖
could represent a drought season and state 𝑗 could represent
rainy season. In the labour market state 𝑖 could represent
being unemployed and state 𝑗 could represent being
employed.
• We know that for a farmer there is probability that a
particular drought year could be followed by a rainy season
and to those in the labour market there is a chance that if
they are unemployed in one period there is certainly chance
they may get a job in the next period.
Mathematics for Economics II Department of Economics 18
• From the above scenarios we can see that the probability of
getting into a particular state 𝑖 whilst coming from state 𝑗 is
essential for both policy and decision making.
• A stochastic matrix is defined as a non-negative square
matrix whose columns add to one.
• It is a matrix of probability or stochastic movements.
• The element 𝑎𝑖𝑗 in the stochastic matrix gives the probability
of having state 𝑖 given the previous period was in state 𝑗.

Mathematics for Economics II Department of Economics 19

• Consider the vector of employment status below
1, employed
𝑋=ቊ
2, unemployed
• Then the stochastic matrix is of the form
𝑎11 𝑎12
𝑃= 𝑎
21 𝑎22
𝑎11 =probability that someone who had a job in period 1
(𝑗 = 1)will have a job in period 2.
𝑎21 =probability that someone who had a job in period 1
(𝑗 = 1)will have employment in period 2.
Mathematics for Economics II Department of Economics 20
 In-class exercise 1.4
Suppose there are 100,000 individuals participating in a
particular labour market. If an individual is employed in the
current period, there is a probability of 0.9 that the person will
be employed in the next period. If the individual is
unemployed in the current period, there is a probability of 0.4
they will have found employment in the next period.
a) Write out the stochastic matrix.
b) What is the equilibrium rate of employment in the market?

Mathematics for Economics II Department of Economics 21

• For a column stochastic, columns add upto one while a row
stochastic, rows add upto one.
Example
0.9 0.5
𝑃=
0.1 0.5
A rainy year is 90% likely to be followed by a rainy year and a
drought year is 50% likely to be followed by another drought
year. Suppose the rainfall condition is known for the initial
(0) 1
year to be ‘rainy’. Then the vector 𝑥 = gives
0
probabilities of rainy and drought for known initial year.
Mathematics for Economics II Department of Economics 22
• For the following year, focused have to be made based on
current condition and the transition matrix.
𝑥 (1) = 𝑃𝑥 (0)
0.9 0.5 1 0.9
= =
0.1 0.5 0 0.1
𝑥 (2) = 𝑃𝑥 (1)
0.9 0.5 0.9 0.86
= =
0.1 0.5 0.1 0.14
⇒ 𝑥 (2) = 𝑃2 𝑥 (0)
Mathematics for Economics II Department of Economics 23
• Probability that year 2 will be rainy is 86% and chances of a
drought is 14%
• In general, with the latest period known with certainty taken
as the initial period, the focus for the 𝑛𝑡ℎ period made in the
initial period is given by
𝑥 (𝑛) = 𝑃𝑥 (𝑛−1) = 𝑃𝑛 𝑥 (0)

Mathematics for Economics II Department of Economics 24

 End of Lecture 1

Mathematics for Economics II Department of Economics 25