ECON 331 Lecture Notes: Ch 4 and Ch 5

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
Matrix of Coefficients Vector of Unknows Vector of Constants
In general
a11 x1 + a12x2 + . . . + a1nxn = d1
a21 x1 + a22x2 + . . . + a2nxn = d2
..............................
am1 x1 + am2x2 + . . . + amnxn = dm
n-unknowns (x1, x2, . . . xn)
Matrix form     
a11 a12 . . . a1n x1 d1

 a21 a22 . . . a2n 
 x2  
  d2 

 .. ... ..  .. = .. 
 . ... .  .   . 
am1 am2 . . . amn xn dm
Matrix shorthand
Ax = d
Where:
A= coefficient martrix or an array
x= vector of unknowns or an array
d= vector of constants or an array
Subscript notation
aij
is the coefficient found in the i-th row (i=1,. . . ,m) and the j-th column (j=1,. . . ,n)

1
1.1 Vectors as special matrices

The number of rows and the number of columns define the DIMENSION 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
x = [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 iff 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
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
 
ka11 ka12 . . . ka1n
 ka21 ka22 . . . ka2n 
kA =  .. 

. 
kam1 kam2 . . . kamn
Example

Let k=[3] and A= 64 25
then kA=
× 6 3 × 2 = 18 6
kA = 33 ×



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.

2
Example
A × B = C
(1 × 3) (3 × 4) (1 × 4)
1row 3rows 1row
3cols 4cols 4cols
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 + a12b21 (sum of row 1 times column 1)
c12=a11 b12 + a12b22 (sum of row 1 times column 2)
c21=a21 b11 + a22b21 (sum of row 2 times column 1)
c22=a21 b12 + a22b22 (sum of row 2 times column 2)
Example 2


1 2

3 2 3 4 =

(3 × 1) +(2 × 3) (3 × 2) +(2 × 4)

= 9 14

Example 3  
2
3 2 1  1  = (3 × 2) + (2 × 1) + (1 × 4) = [12]
 
4
12 is the inner product of two vectors.
Suppose

x = xx1

then x = x1 x2
2

therefore

x x = x1 x2
 x1
x2

= x21 + x22
However
xx = 2 by 2 matrix



x1 
x1 x2 = xxx1 xx1x2 2
2

x2 2 1 2

Example 4  
1 3
A= 2 8
4 0


b= 5
9
(1 × 5) + (3 × 9)
   
32
Ab =  (2 × 5) + (8 × 9) = 82 
(4 × 5) + (0 × 9) 20

3
 Example
Ax = d

 A   x  d 
6 3 1 x1 22
 1 4 −2   x2  =  12 
4 −1 5 x3 10
(3 × 3) (3 × 1) (3 × 1)
This produces
6x1 + 3x2 + x3 = 22
x1 + 4x2 − 2x3 = 12
4x1 − x2 + 5x3 = 10
1.1.1 National Income Model
y = c + I0 + G 0
C = a + bY
Arrange as
y − C = I0 + G0
−bY + C = a
Matrix form
A x =d
1 −1
Y


I
= o +a Go


−b 1 C
1.1.2 Division in Matrix Algebra
In ordinary algebra
a =c
b
is well defined iff b
= 0.
Now 1b can be rewritten as b 1 , therefore ab 1 = c, also b 1a = c.
− − −

But in matrix algebra

A
B =C
is not defined. However,
AB 1 = C−

is well defined. BUT

AB 1
= B 1 A
− −

B 1 is called the inverse of B

−

B 1
= B1
−

In some ways B −1
has the same properties as b −1
but in other ways it differs. We will explore these
differences later.

4
1.2 Linear Dependance

Suppose we have two equations

x1 + 2x2 = 1
3x1 + 6x2 = 3
To solve
3 [−2x2 + 1] − 6x2 = 3
6x2 + 3 − 6x2 = 3
3=3
There is no solution. These two equations are linearly dependent. Equation 2 is equal to two times
equation one.



1 2 x1 = 1
3 6 x 3 2

Ax = d
where A is a two column vectors   
U1 U2
 1  2 
3 6
Or A is two row vector
V1 = [1 2]
V2 = [3 6]
Where column two is twice column one and/or row two is three times row one
2U1 = U2 or 3V1 = V2
Linear Dependence Generally:
A set of vectors is said to be linearly dependent iff any one of them can be expressed as a linear combination
of the remaining vectors.
Example:
Three vectors,



V1 = 27 V2 = 18 V3 = 45
are linearly dependent since
3V1 −
2V2 = V3


6 2 4
21 − 16 = 5
or expressed as
3V1 − 2V2 − V3 = 0
General Rule
A set of vectors, V1, V2,..., Vn are linearly dependent if there exsists a set of scalars (i=1,. . . ,n). Not all
equal to zero, such that
n

= ki Vi = 0
i=1
Note n

kiVi = k1V1 + k2V2 + . . . + knVn

i=1

5
1.3 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.

1.3.1 Communicative Law of Addition

A+B =B+A
Since we are adding individual elements and aij + bij = bij + aij for all i and j.

1.3.2 Similarly Associative Law of Addition

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

1.3.3 Matrix Multiplication

Matrix multiplication in not communtative
AB
= 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
Let A= 13 24 and B= 06 −71

AB = (1(3××10) + (2 × 6) (1 × −1) + (2 × 7)



12 13
0) + (4 × 6) (3 × −1) + (4 × 7) = 24 25
But
BA = (0)(1) − (1)(3) (0)(2) − (1)(4)

−3 −4

(6)(1) + (7)(3) (6)(2) + (7)(4) = 27 40
Therefore, we realize the distinction of post multiply and pre multiply. In the case
AB = C
B is pre multiplied by A, A is post multiplied by B.

6
1.3.4 Associative Law
Matrix multiplication is associative
(AB)C = A(BC ) = ABC
as long as their dimensions conform to our earlier rules of multiplication.
A × B × C
(m × n ) (n × p) (p × q)
1.3.5 Distributive Law
Matrix multiplication is distributive
A(B + C ) = AB + AC Pre multiplication
(B + C )A = BA + CA Post multiplication
1.4 Identity Matrices and Null Matrices

1.4.1 Identity matrix:

is a square matrix with ones on its principal diagonals and zeros everywhere else.
 
  1 0 ... n

1 0 0  .. 
I2 = 10 01 0 1 .
 
I3 =  0 1 0  In = 
 .. .


0 0 1  . . . 0 
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

Let A= 12 34

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





IA = 10 01 2 4 (0 × 1) + (1 × 2) (0 × 3) + (1 × 4) 2 4

Example 2

Let A= 12 20 33




IA = 10 01 1 2 3 = 1 2 3 = A {I Case}
2 0 3 2 0 3 2

 

1 0 0

AI = 12 20 33  0 1 0 = 1 2 3 = A {I Case}
2 0 3 3
0 0 1
Furthermore,
AIB = (AI )B = A(IB) = AB
(m × n)(n × p) (m × n)(n × p)

7
1.4.2 Null Matrices
A null matrix is simply a matrix where all elements equal zero.



0 = 00 00 0 = 00 00 00
(2 × 2) (2 × 3)
The rules of scalar algebra apply to matrix algebra in this case.
Example
a + 0 = a ⇒ {scalar}



A + 0 = aa11 aa12 + 00 00 = A {matrix}
21 22
 
0


A × 0 = aa11 aa12 aa13 =  0  = 00 = 0
21 22 23
0
1.5 Idiosyncracies of matrix algebra
1) We know AB
=BA
2)ab=0 implies a or b=0
In matrix
AB = 21 42


−24 0 0



1 −2 = 0 0
1.5.1 Transposes and Inverses
1)Transpose: is when the rows and columns are interchanged.
Transpose of A=A’or AT
Example
If A= 31 80 −49 and B= 31 47

 
3 1

A’= 8 0  and B’= 31 47
−9 4
Symmetrix Matrix
   
1 0 4 1 0 4
If A= 0 3 7  then A’= 0 3 7 
4 7 2 4 7 2
A is a symmetric matrix.

Properties of Transposes 1) (A’)’=A

2) (A+B)’=A’+B’
3) (AB)’=B’A’
Inverses and their Properties
In scalar algebra if
ax = b

8
 then
x = ab or ba −1

In matrix algebra
if
Ax = d
then
x = A 1d −

where A 1 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 sifficient)
3) In scalar algebra aa = 1, in matrix algebra AA 1 = A 1A = I − −

4) If an inverse exists then it must be unique.

Example
3 1

1
Let A= 0 2 and A 1 = 03 −
−
1
1
6

2 −1 by factoring  1 is a scalar


A =
−1 1
0 3
6 6

Post Multiplication
2 −1 1 = 6 0 1 = 1 0





AA 1 = 30 12
−

0 3 6 0 6 6 0 1
Pre Multiplication
A A = 16 02 −31 30 12 = 16 60 06 = 10 01





−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) (A’) 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
Ax = d

9
 1  −1813 −2616 −1310
       
6 3 1 x1 22
A= 1 4 −2  x =  x2  d= 12  A 1 = 52
− 
4 −1 5 x3 10 −17 18 21
then
18 −16 −10
      
x1 1 22 2
 x2 = −13 26 13  12 = 3 
x3 52 −17 18 21 10 1
x1 = 2 x2 = 3 x3 = 1
∗ ∗ ∗

2 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 coefficient matrix is
|A| = (1)(4) − (2)(2) = 0
a determinant of zero tells us that the equations are linearly dependent. Sometimes called a ”vanishing
determinant.”
In general, the determinant of a square matrix, A is written as |A| or detA.
For two by two case
|A | = a11 a12 = a a − a a = k
a21 a22 11 22 12 21

where k is unique
any k
= 0 implies linear independence
Example 1

A= 31 25
|A| = (3 × 5) − (1 × 2) = 13 {Non-singular}
Example 2

B= 28 246
|B| = (2 × 24) − (6 × 8) = 0 {Singular}
Three by three case

10
  
a1 a2 a3
Given A= b1 b2 b3 
c1 c2 c3
then
|A| = (a b c ) + (a b c ) + (b c a ) − (a b c ) − (a b c ) − (b c a )
1 2 3 2 3 1 1 2 3 3 2 1 2 1 3 3 2 1

Cross-diagonals  
a1 a2 a3
 b1 b2 b3 
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.
2.1 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
 or Minors

a1 a2 a3
Given A= b1 b2 b3 
c1 c2 c3
By deleting the first row and first column, we get


|M11| = cb22 cb33
The determinant of this matrix is the minor element a1 .
|Mij | ≡ is the subdeterminant from deleting the i-th row and the j-th column.
 
a1 a2 a3
Given A= b1 b2 b3 
c1 c2 c3
then


M21 ≡ aa12 aa13 M31≡
a12 a13
a22 a23
32 33

2.1.1 Cofactors
A cofactor is a minor with a specific algebraic sign.
Cij = (−1)i+j |Mij |
therefore
C11 = (−1)2 |M11 | = |M11|

C21 = (−1)3 |M21| = −|M21 |

The determinant by Laplace
Expanding down the first column  
a11 a12 a13
A =  a21 a22 a23 
a31 a32 a33

11
 |A| = a |C | + a |C | + a |C | =
ai |Ci |
3

11 11 21 21 31 31 1 1
i=1
     
|A| 
= a11  a22 a23 
− 
a21  a12 a13   a12 a13 
a32 a33  + a31  a22 a23
  
a32 a33  

Note: minus sign (-1)(1+2)

|A| = a[a22 a33 − a23 a32 ] − a21 [a12 a33 − a13 a32] + a31 [a12 a23 − a13 a22]
11

Laplace expansion can be used to expand along any row or any column.
Example
Third row      
|A| = a 
 a12 a13 
 −a 
 a11 a13  + a  a11 a12 

31 
a22 a23  32 
a21 a23  33  a21 a22 
Example
 
8 1 3
A= 4 0 1 
6 0 3
(1)Expand the first column
     
|A | 
= 8  0 1 
 − 
4  1 3  + 6  1 3 

0 3  0 3 0 1 

|A| = (8 × 0) − (4 × 3) + (6 × 1) = −6
(2)Expand the second column
     
|A| = − 
1  4 1  + 0  8 3 
 − 
0  8 3 

6 3 6 3  4 1 

|A| = (−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.)

2.2 Rank of a Matrix

Definition
The rank of a matrix is the maximum number linearly independent rows in the matrix.
If A is an m×n matrix, then the rank of A is
r(A) ≤ min [m, n]
Read as: the rank of A is less than or equal to the minimum of m or n.
Using Determinants to Find the Rank
(1) If A is n×m and |A|=0
(2) Then delete one row and one column, and find the determinant of this new (n-1)×(n-1) matrix.
(3) Continue this process until you have a non-zero determinant.

12
3 Matrix Inversion
Given an n×n matrix, A, the inverse of A is
A 1 = |A1 | • AdjA
−

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 cofactors of the elements of A
If A = aa11 aa12


then C = | C | |C |
= a − a

| C | |C | −a a
11 12 22 21
21 22 21 22 12 11

Example

3 2
Let A= 1 0 ⇒ |A|=-2
Step 1: Find the cofactor matrix
C= |C | |C |
= 0 − 1

|C | |C | −2 3
11 12
21 22

Step 2: Transpose the cofactor matrix

0 −2


C T = AdjA = −1 3
Step 3: Multiply all the elements of AdjA by A1 to find A 1
| |
−

A 1 = |A1 | • AdjA = − 12 −01 −32 = 01 −13

 


−

2 2

Step 4: Check by AA = I
−1

0 1 = (3)(0) + (2)( 12 ) (3)(1) + (2)(− 32 )

3 2
1 0
12 − 32 (1)(0) + (0)( 12 ) (1)(1) + (0)(− 32 )
= 10 01

4 Cramer’s Rule
Suppose:
Equation 1 a1 x1 + a2x2 = d1

Equation 2 b1 x1 + b2 x2 = d2
or
A
x
= d

a1 a2 x1 = d1
b1 b2 x2 d2
where
A = a1b2 − a2b1
= 0
Solve for x1 by substitution
From equation 1
x2 = d1 −a a1x1
2

13
 and equation 2
x2 = d2 −b b1x1
2

therefore:
d1 − a1x1 = d2 − b1x1
a2 b2
Cross multiply
d1b2 − a1b2x1 = d2 a2 − b1a2x1
Collect terms
d1 b2 − d2a2 = (a1 b2 − b1a2)x1

x1 = da1bb2 −− db 2aa2
1 2 1 2

The denominator is the determinant of |A|

The numerator is the same as the denominator except d1 d2 replaces a1b1.
Cramer’s Rule
 
d1 
 a2 
d  b2  d1b2 − d2a2
x1 = a2 
 a2  = a1b2 − b1a2
1 
b1  b2 
Where the d vector replaces column 1 in the A matrix
To find x2 replace column 2 with the d vector
 
a1  d1 
d2  a1d2 − d1b1

b 
x2 = a1 
 a2  = a1b2 − b1a2
1 
b1  b2 
Generally: to find xi,replace column i with vector d; find the determinant.
xi = the ratio of two determinants
xi = AA
|
|
i|
|

4.0.1 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

| A
| = (1)(1) − (−1)(1) = 2

Find Qe  
 10 1 
 
 
Qe =
2 1 10 − 2
= =4
2 2

14
 Find Pe 



 1 10 
 
Pe = −1

2
2
=
2 − (−10)
2
=6

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

Qd = 10 P −

10 − 6 = 4

4.0.2 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 Ye 
 I0 + G 0

 1  −

Y e =  a1 1 1 = I0 +1 G0b+ a
 
  − −
 b 1 −

Solve for Ce  

 1 I0 + G0 

Ce =
 −

b a  
=
a + b(I0 + G0)

 1 −1 
 1− b
 − b 1 

15