Lecture Notes for Chapters 4 & 5

Kevin Wainwright
April 26, 2014

1 Matrix Algebra
1. Gives us a shorthand way of writing a large system of equations.
2. Allows us to test for the existance of solutions to simultaneous
systems.
3. Allows us to solve a simultaneous system.
DRAWBACK: Only works for linear systems. However, we can
often covert non-linear to linear systems.
Example
y = axb
ln y = ln a + b ln x
Matrices and Vectors
Given
y = 10 x ) x + y = 10
y = 2 + 3x ) 3x + y = 2
In matrix form
1 1 x 10
=
3 1 y 2

1
 Matrix of Coe¢ cients Vector of Unknows Vector of Constants

In general

a11 x1 + a12 x2 + : : : + a1n xn = d1

a21 x1 + a22 x2 + : : : + a2n xn = d2
::::::::::::::::::::::::::::::
am1 x1 + am2 x2 + : : : + amn xn = dm

n-unknowns (x1; x2; : : : xn )

Matrix form
2 32 3 2 3
a11 a12 : : : a1n x1 d1
6 a21 a22 : : : a2n 7 6 x2 7 6 d2 7
6 . . . . ... 7 6 .. 7 = 6 . 7
4 .. ::: 54 . 5 4 .. 5
am1 am2 : : : amn xn dm

Matrix shorthand
Ax = d
Where:
A= coe¢ cient martrix or an array
x= vector of unknowns or an array
d= vector of constants or an array

Subscript notation
aij
is the coe¢ cient found in the i-th row (i=1,. . . ,m) and the j-th
column (j=1,. . . ,n)

2
1.1 Vectors as special matrices

The number of rows and the number of columns de…ne the DIMEN-
SION of a matrix.
A is m rows and n is columns or ”mxn.”
A matrix containing 1 column is called a ”column VECTOR”
x is a n 1 column vector
d is a m 1 column vector
If x were arranged in a horizontal array we would have a row vector.
Row vectors are denoted by a prime

x0 = [x1 ; x2 ; : : : ; xn ]

A 1 1 vector is known as a scalar.

x = [4] is a scalar

Matrix Operators
If we have two matrices, A and B, then

A = B if f aij = bij
Addition and Subtraction of Matrices
Suppose A is an m n matrix and B is a p q matrix then A and
B is possible only if m=p and n=q. Matrices must have the same
dimensions.
a11 a12 b11 b12 (a11 + b11 ) (a12 + b12 ) c11 c12
+ = =
a21 a22 b21 b22 (a21 + b21 ) (a22 + b22 ) c21 c22

Subtraction is identical to addition

9 4 7 2 (9 7) (4 2) 2 2
= =
3 1 1 6 (3 1) (1 6) 2 5

3
 Scalar Multiplication
Suppose we want to multiply a matrix by a scalar
k A
1 1 m n
We multiply every element in A by the scalar k
2 3
ka11 ka12 : : : ka1n
6 ka21 ka22 : : : ka2n 7
kA = 64 ...
7
5
kam1 kam2 : : : kamn
Example
6 2
Let k=[3] and A=
4 5
then kA=
3 6 3 2 18 6
kA = =
3 4 3 5 12 15
Multiplication of Matrices
To multiply two matrices, A and B, together it must be true that
for
A B = C
m n n q m q
That A must have the same number of columns (n) as B has rows
(n).
The product matrix, C, will have the same number of rows as A
and the same number of columns as B.
Example
A B = C
(1 3) (3 4) (1 4)
1row 3rows 1row
3cols 4cols 4cols
4
 In general
A B C D = E
(3 2) (2 5) (5 4) (4 1) (3 1)
To multiply two matrices:
(1) Multiply each element in a given row by each element in a given
column
(2) Sum up their products
Example 1
a11 a12 b11 b12 c11 c12
=
a21 a22 b21 b22 c21 c22
Where:
c11 =a11 b11 + a12 b21 (sum of row 1 times column 1)
c12 =a11 b12 + a12 b22 (sum of row 1 times column 2)
c21 =a21 b11 + a22 b21 (sum of row 2 times column 1)
c22 =a21 b12 + a22 b22 (sum of row 2 times column 2)
Example 2
1 2
3 2 = (3 1) +(2 3) (3 2) +(2 4) = 9 14
3 4
Example 3
3 2
2
3 2 1 415= (3 2) + (2 1) + (1 4) = [12]
4
12 is the inner product of two vectors.
Suppose
x1
x= then x0 = x1 x2
x2
5
therefore
x1
x0 x = x1 x2
x2

= x21 + x22
However
xx0 = 2 by 2 matrix

x1 x21 x1 x2
x1 x2 =
x2 x2 x1 x22
Example 4 2
3
1 3
A=42 85
4 0
5
b=
9
2 3 2 3
(1 5) + (3 9) 32
Ab = 4 (2 5) + (8 9) 5 = 4 82 5
(4 5) + (0 9) 20
Example
Ax = d

2 A 3 2 x3 2 d 3
6 3 1 x1 22
41 4 2 5 4 5
x2 = 4 12 5
4 1 5 x3 10
(3 3) (3 1) (3 1)

6
 This produces
6x1 + 3x2 + x3 = 22
x1 + 4x2 2x3 = 12
4x1 x2 + 5x3 = 10

1.1.1 National Income Model

y = c + I0 + G0
C = a + bY
Arrange as
y C = I0 + G0
bY + C = a
Matrix form
A x =d
1 1 Y Io + Go
=
b 1 C a

1.1.2 Division in Matrix Algebra

In ordinary algebra
a
=c
b
is well de…ned i¤ b6= 0.
Now 1b can be rewritten as b 1 ; therefore ab 1
= c; also b 1 a = c:

But in matrix algebra

A
=C
B
is not de…ned. However,
1
AB =C

7
 is well de…ned. BUT
1
AB 6= B 1 A
1
B is called the inverse of B
1 1
B 6=
B
In some ways B 1 has the same properties as b 1
but in other ways
it di¤ers. We will explore these di¤erences later.

1.2 Commutative, Associative, and Distributive

Laws

From Highschool algebra we know commutative law of addition,

a+b=b+a
commutative law of multiplication,

ab = ba

Associative law of addition,

(a + b) + c = a + (b + c)

associative law of multiplication,

(ab)c = a(bc)

Distributive law
a(b + c) = ab + ac
In matrix algebra most, but not all, of these laws are true.

8
1.2.1 I) Communicative Law of Addition
A+B =B+A
Since we are adding individual elements and aij + bij = bij + aij for
all i and j.

1.2.2 II) Similarly Associative Law of Addition

A + (B + C) = (A + B) + C
for the same reasons.

1.2.3 III) Matrix Multiplication

Matrix multiplication in not communtative

IB 6= BA

Example 1
Let A be 2 3 and B be 3 2
A B = C whereas B A = C
(2 3) (3 2) (2 2) (3 2) (2 3) (3 3)

Example 2
1 2 0 1
Let A= and B=
3 4 6 7

(1 10) + (2 6) (1 1) + (2 7) 12 13
AB = =
(3 0) + (4 6) (3 1) + (4 7) 24 25

9
 But
(0)(1) (1)(3) (0)(2) (1)(4) 3 4
BA = =
(6)(1) + (7)(3) (6)(2) + (7)(4) 27 40

Therefore, we realize the distinction of post multiply and pre mul-

tiply. In the case
AB = C
B is pre multiplied by A, A is post multiplied by B.

1.2.4 IV) Associative Law

Matrix multiplication is associative

(AB)C = A(BC) = ABC

as long as their dimensions conform to our earlier rules of multipli-

cation.
A B C
(m n) (n p) (p q)

1.2.5 V) Distributive Law

Matrix multiplication is distributive

A(B + C) = AB + AC Pre multiplication

(B + C)A = BA + CA Post multiplication

10
1.3 Identity Matrices and Null Matrices

1.3.1 Identity matrix:

is a square matrix with ones on its principal diagonals and zeros every-
where else.
2 3
2 3 1 0 ::: n
1 0 0 6 .. 7
1 0 4 5 60 1 . 7
In = 6 .
.. 0 7
I2 = I3 = 0 1 0 .
0 1 4 .. 5
0 0 1
0 ::: 0 1

Identity Matrix in scalar algebra we know

1 a=a 1=a

In matrix algebra the identity matrix plays the same role

IA = AI = A

Example 1
1 3
Let A=
2 4

1 0 1 3 (1 1) + (0 2) (1 3) + (0 4) 1 3
IA = = =
0 1 2 4 (0 1) + (1 2) (0 3) + (1 4) 2 4

Example 2
1 2 3
Let A=
2 0 3

1 0 1 2 3 1 2 3
IA = = = A fI2 Caseg
0 1 2 0 3 2 0 3
11
 2 3
1 0 0
1 2 3 40 1 05= 1 2 3
AI = = A fI3 Caseg
2 0 3 2 0 3
0 0 1

Furthermore,

AIB = (AI)B = A(IB) = AB

(m n)(n p) (m n)(n p)

1.3.2 Null Matrices

A null matrix is simply a matrix where all elements equal zero.

0 0 0 0 0
0= 0=
0 0 0 0 0
(2 2) (2 3)

The rules of scalar algebra apply to matrix algebra in this case.

Example
a + 0 = a ) fscalarg

a11 a12 0 0
A+0= + =A fmatrixg
a21 a22 0 0
23
0
a11 a12 a13 0
A 0= =405= =0
a21 a22 a23 0
0

12
1.4 Idiosyncracies of matrix algebra
1) We know AB6=BA
2)ab=0 implies a or b=0
In matrix
2 4 2 4 0 0
AB = =
1 2 1 2 0 0

1.4.1 Transposes and Inverses

1)Transpose: is when the rows and columns are interchanged.
Transpose of A=A’or AT

Example
3 8 9 3 4
If A= and B=
1 0 4 1 7
2 3
3 1
3 4
A’=4 8 0 5 and B’=
1 7
9 4
Symmetrix Matrix
2 3 2 3
1 0 4 1 0 4
If A=4 0 3 7 5 then A’=4 0 3 75
4 7 2 4 7 2
A is a symmetric matrix.

Properties of Transposes
1) (A0)0 = A
2) (A + B)0 = A0 + B0
3) (AB)0 = B0A0

Inverses and their Properties

13
 In scalar algebra if
ax = b
then
b 1
x= or ba
a
In matrix algebra, if
Ax = d
then
x = A 1d
1
where A is the inverse of A.

Properties of Inverses
1) Not all matrices have inverses
non-singular: if there is an inverse
singular: if there is no inverse
2) A matrix must be square in order to have an inverse. (Necessary
but not si¢ cient)
3) In scalar algebra aa = 1; in matrix algebra AA 1 = A 1 A = I
4) If an inverse exists then it must be unique.

Example
1 1
3 1 1 3 6
Let A= and A = 1
0 2 0 2

1 2 1
1 1
A = 6 by factoring 6 is a scalar
0 3
Post Multiplication

1 3 1 2 1 1 6 0 1 1 0
AA = = =
0 2 0 3 6 0 6 6 0 1

14
 Pre Multiplication
1 2 1 3 1 1 6 0 1 0
A 1A = = =
6 0 3 0 2 6 0 6 0 1
Further properties
If A and B are square and non-singular then:
1) (A 1 ) 1 = A
2) (AB) 1 = B 1 A 1
3) (A0) 1 = (A 1 )1
Solving a linear system
Suppose
A x = d
(3 3) (3 1) (3 1)
then
A 1 A x = A 1 d
(3 3) (3 3) (3 1) (3 3) (3 1)
I x = A 1 d
(3 3) (3 1) (3 3) (3 1)
x = A 1d
Example
2 3 2 Ax
3 =d 2 3 2 3
6 3 1 x1 22 18 16 10
1 4
A=41 4 2 5 x = 4 x2 5 d = 4 12 5 A 1
= 13 26 13 5
4 1 5 x3 10 52 17 18 21
then 2 3 2 32 3 2 3
x1 18 16 10 22 2
4 x2 5 = 1 4 13 26 13 5 4 12 5 = 4 3 5
x3 52 17 18 21 10 1
x1 = 2 x2 = 3 x3 = 1

15
1.5 Linear Dependence and Determinants

Suppose we have the following

1: x1 + 2x2 = 1
2: 2x1 + 4x2 = 2
where equation two is twice equation one. Therefore, there is no
solution for x1 ; x2 :
In matrix form:
Ax = d

" A #" x # " d #

1 2 x1 1
=
2 4 x2 2
The determinant of the coe¢ cient matrix is

jAj = (1)(4) (2)(2) = 0

a determinant of zero tells us that the equations are linearly de-

pendent. Sometimes called a ”vanishing determinant.”
In general, the determinant of a square matrix, A is written as jAj
or detA.
For two by two case

a11 a12
jAj = = a11 a22 a12 a21 = k
a21 a22
where k is unique
any k6= 0 implies linear independence
Example 1

16
 3 2
A=
1 5

jAj = (3 5) (1 2) = 13 fNon-singularg

Example 2
2 6
B=
8 24

jBj = (2 24) (6 8) = 0 fSingularg

Three by three
2 case 3
a1 a2 a3
4
Given A= b1 b2 b3 5
c1 c2 c3
then

jAj = (a1 b2 c3 ) + (a2 b3 c1 ) + (b1 c2 a3 ) (a3 b2 c1 ) (a2 b1 c3 ) (b3 c2 a1 )

Cross-diagonals 2 3
a1 a2 a3
4 b1 b2 b3 5
c1 c2 c3
Use viso to display cross diagonals

Multiple along the diagonals and add up their products

)The product along the BLUE lines are given a positive sign
)The product of the RED lines are negative.

17
1.6 Using Laplace expansion
) The cross diagonal method does not work for matrices greater than
three by three
) Laplace expansion evaluates the determinant of a matrix, A, by
means of subdeterminants of A.
Subdeterminants
2 or Minors
3
a1 a2 a3
Given A=4 b1 b2 b3 5
c1 c2 c3
By deleting the …rst row and …rst column, we get
b2 b3
jM11 j =
c2 c3
The determinant of this matrix is the minor element a1 .
jMij j is the subdeterminant from deleting the i-th row and the
j-th column.
2 3
a1 a2 a3
Given A=4 b1 b2 b3 5
c1 c2 c3
then
a12 a13 a12 a13
M21 M31
a32 a33 a22 a23

1.6.1 Cofactors
A cofactor is a minor with a speci…c algebraic sign.
Cij = ( 1)i+j jMij j
therefore
C11 = ( 1)2 jM11 j = jM11 j
C21 = ( 1)3 jM21 j = jM21 j

18
The determinant by Laplace Expanding down the …rst column
2 3
a11 a12 a13
A = 4 a21 a22 a23 5
a31 a32 a33
3
X
jAj = a11 jC11 j + a21 jC21 j + a31 jC31 j = ai1 jCi1 j
i=1

a22 a23 a12 a13 a a

jAj = a11 a21 + a31 12 13
a32 a33 a32 a33 a22 a23
Note: minus sign ( 1)(1+2)

jAj = a11 [a22 a33 a23 a32 ] a21 [a12 a33 a13 a32 ] + a31 [a12 a23 a13 a22 ]

Laplace expansion can be used to expand along any row or any

column.
Example: Third row

a12 a13 a11 a13 a a

jAj = a31 a32 + a33 11 12
a22 a23 a21 a23 a21 a22

Example
2 3
8 1 3
A=44 0 15
6 0 3
(1)Expand the …rst column

0 1 1 3 1 3
jAj = 8 4 +6
0 3 0 3 0 1

19
 jAj = (8 0) (4 3) + (6 1) = 6
(2) Expand the second column

4 1 8 3 8 3
jAj = 1 +0 0
6 3 6 3 4 1

jAj = ( 1 6) + (0) (0) = 6

Suggestion: Try to choose an easy row or column to expand. (i.e. the
ones with zero’s in it.)

1.7 Matrix Inversion

Given an n n matrix, A, the inverse of A is

1 1
A = AdjA
jAj
where AdjA is the adjoint matrix of A. AdjA is the transpose of
matrix A’s cofactor matrix. It is also the adjoint, which is an n n
matrix

Cofactor Matrix (denoted C)

The cofactor matrix of A is a matrix who’s elements are the cofac-
tors of the elements of A

a11 a12 jC11 j jC12 j a22 a21

If A = then C = =
a21 a22 jC21 j jC22 j a12 a11

Example
3 2
Let A= ) jAj=-2
1 0

20
 Step 1: Find the cofactor matrix

jC11 j jC12 j 0 1
C= =
jC21 j jC22 j 2 3

Step 2: Transpose the cofactor matrix

0 2
C T = AdjA =
1 3
1 1
Step 3: Multiply all the elements of AdjA by jAj to …nd A

1 1 1 0 2 0 1
A = AdjA = = 1 3
jAj 2 1 3 2 2

1
Step 4: Check by AA =I

3 2 0 1 (3)(0) + (2)( 12 ) (3)(1) + (2)( 3

2)
1 3 =
1 0 2 2 (1)(0) + (0)( 12 ) (1)(1) + (0)( 3
2)
1 0
=
0 1

1.8 Cramer’s Rule

Suppose:

Equation 1 a1 x1 + a2 x2 = d1
Equation 2 b1 x1 + b2 x2 = d2
or
A x = d
a1 a2 x1 d1
=
b1 b2 x2 d2

21
where
A = a1 b 2 a2 b1 6= 0
Solve for x1 by substitution
From equation 1
d1 a1 x1
x2 =
a2
and equation 2
d2 b1 x1
x2 =
b2
therefore:
d1 a1 x 1 d2 b1 x1
=
a2 b2
Cross multiply
d1 b2 a1 b2 x1 = d2 a2 b 1 a2 x 1
Collect terms
d1 b2 d2 a2 = (a1 b2 b1 a2 )x1
d1 b2 d2 a2
x1 =
a1 b2 b1 a2
The denominator is the determinant of jAj and the numerator is the
same as the denominator except d1 d2 replaces a1 b1 :

Cramer’s Rule

d1 a2
d2 b2 d1 b2 d 2 a2
x1 = =
a1 a2 a1 b2 b 1 a2
b1 b2
Where the d vector replaces column 1 in the A matrix

22
 To …nd x2 replace column 2 with the d vector

a1 d1
b1 d2 a1 d2 d1 b1
x2 = =
a1 a2 a1 b2 b 1 a2
b1 b2

Generally: to …nd xi ;replace column i with vector d; …nd the de-

terminant.
xi = the ratio of two determinants
xi = jA ij
jAj

1.8.1 Example: The Market Model

Equation 1 Qd = 10 P Or Q + P = 10
Equation 2 Qs = P 2 Or Q+P =2
Matrix form
A x = d
1 1 Q 10
=
1 1 P 2
jAj = (1)(1) ( 1)(1) = 2
Find Qe
10 1
2 1 10 2
Qe = = =4
2 2
Find Pe
1 10
1 2 2 ( 10)
Pe = = =6
2 2

23
Substitute P and Q into either equation 1 or equation 2 to verify

Qd = 10 P
10 6 = 4

1.8.2 Example: National Income Model

Y = C + I0 + G0 Or Y C = I0 + G0
C = a + bY Or bY + c = a
In matrix form
1 1 Y I0 + G0
=
b 1 C a

Solve for Y e
I0 + G0 1
a 1 I0 + G0 + a
Ye = =
1 1 1 b
b 1
Solve for C e
1 I0 + G0
b a a + b(I0 + G0 )
Ce = =
1 1 1 b
b 1

Numeric example:
Let C = 100 + 0:75Y; I = 150 and G = 250. Then the model is

Y C = I +G
Y C = 400

24
and

C = 100 + 0:75Y
0:75Y C = 100

In Matrix form
1 1 Y 400
=
0:75 1 C 100

Solve for Y e
400 1
100 1 500
Ye = = = 2000
1 1 0:25
0:75 1
Solve for C e
1 400
0:75 100 100 + 0:75(400)
Ce = = = 1600
1 1 0:25
0:75 1

25