Lecture 3

Cramer’s rule and Inverse matrix method

Sections 5.5, 4.6 from
Fundamental methods of Mathematical Economics, McGraw Hill 2005, 4th Edition.
By A. C. Chiang & Kevin Wainwright are covered.

Non-homogenous systems of linear equations:

An arbitrary system of m linear equations in n unknowns of form

𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑑1

𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑑2

𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑑𝑚

Where x1,x2,...xn are unknowns and a’s and d’s denote constants, is known as non-
homogeneous system provided 𝑑’𝑠 𝑛𝑜𝑡 𝑎𝑙𝑙 𝑧𝑒𝑟𝑜 (at least one of them is non-zero).

This system can be expressed as 𝐴𝑋 = 𝑏

where

𝑎11 𝑎12 … 𝑎1𝑛 𝑥1 𝑑1

𝑎21 𝑎22 … 𝑎2𝑛 𝑥2 𝑑2
𝐴=[ ⋮ ⋮ ⋮ ⋮ ],𝑋 = [ ⋮ ],𝑏 = [ ⋮ ]
𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛 𝑥𝑛 𝑑𝑚

A is called matrix of coefficients.

Cramer’s rule: If AX=b is a system of n linear equations in n unknowns such that det(A)≠0,
then the system has a unique solution. This solution is

|𝐴1 | |𝐴2 | |𝐴𝑛 |

𝑥1 = , 𝑥2 = , … , 𝑥𝑛 =
|𝐴| |𝐴| |𝐴|

Where Aj is the matrix obtained by replacing the entries in the jth column of A by the entries
in the matrix

1
 𝑑1
𝑑
𝑏 = [ 2]
⋮
𝑑𝑛

Example 1: Use Cramer’s rule to solve

4𝑥1 + 5𝑥2 = 2

11𝑥1 + 𝑥2 + 2𝑥3 = 3

𝑥1 + 5𝑥2 + 2𝑥3 = 1

Solution:

4 5 0 2
𝐴 = [11 1 2] , 𝑑 = [3]
1 5 2 1

2 5 0 4 2 0 4 5 2
𝐴1 = [3 1 2] , 𝐴2 = [11 3 2] , 𝐴3 = [11 1 3]
1 5 2 1 1 2 1 5 1

|𝐴| = −132, |𝐴1 | = −36, |𝐴2 | = −24, |𝐴3 | = 12

Therefore

|𝐴1 | −36 3
𝑥1 = = =
|𝐴| −132 11

|𝐴2 | −24 2
𝑥2 = = =
|𝐴| −132 11

|𝐴3 | 12 1
𝑥3 = = =−
|𝐴| −132 11

Example 2: Use Cramer’s rule to solve

𝑥1 − 4𝑥2 + 𝑥3 = 6

4𝑥1 − 𝑥2 + 2𝑥3 = −1

2
 2𝑥1 + 2𝑥2 − 3𝑥3 = −20

Solution:

1 −4 1 6
𝐴 = [4 −1 2 ] , 𝑑 = [ −1 ]
2 2 −3 −20

6 −4 1 1 6 1 1 −4 6
𝐴1 = [ −1 −1 2 ] , 𝐴2 = [4 −1 2 ] , 𝐴3 = [4 −1 −1 ]
−20 2 −3 2 −20 −3 2 2 −20

|𝐴| = −55, |𝐴1 | = 144, |𝐴2 | = 61, |𝐴3 | = −230

Therefore

|𝐴1 | 144 |𝐴2 | 61 |𝐴3 | −230 46

𝑥1 = =− , 𝑥2 = = − , 𝑥3 = = =
|𝐴| 55 |𝐴| 55 |𝐴| −55 11

Example 3: Use Cramer’s rule to solve following homogeneous system

𝑥1 + 2𝑥3 = 0

−3𝑥1 + 4𝑥2 + 6𝑥3 = 0

−𝑥1 − 2𝑥2 + 3𝑥3 = 0

Solution:

1 0 2 0
𝐴 = [−3 4 6] , 𝑑 = [0]
−1 −2 3 0

0 0 2 1 0 2 1 0 0
𝐴1 = [0 4 6] , 𝐴2 = [−3 0 6] , 𝐴3 = [−3 4 0]
0 −2 3 −1 0 3 −1 −2 0

|𝐴| = 44, |𝐴1 | = 0, |𝐴2 | = 0, |𝐴3 | = 0

Therefore

3
 |𝐴1 | 0 |𝐴2 | 0 |𝐴3 | 0
𝑥1 = = = 0, 𝑥2 = = = 0, 𝑥3 = = =0
|𝐴| 44 |𝐴| 44 |𝐴| 44

Thus for a homogeneous system if |𝐴| ≠ 0, then only trivial solution exists and we can find
it by Cramer’s rule.

Remark: Cramer’s rule is applicable if A is square matrix and |𝐴| ≠ 0.

Question: Given the system of equations below , solve for z ONLY.

2𝑥 + 𝑦 + 𝑧 = 1
𝑥 – 𝑦 + 4𝑧 = 0
𝑥 + 2𝑦 – 2𝑧 = 3

Inverse matrix method

The application of the concept of inverse matrix to the solution of a simultaneous equation
system is immediate and direct.

A non-homogeneous linear system can be expressed as

𝐴𝑋 = 𝑑

Now if 𝐴−1 exists, the premultiplication of both sides of above equation by 𝐴−1 yields

𝐴−1 𝐴𝑋 = 𝐴−1 𝑑 or 𝑋 = 𝐴−1 𝑑 yields solution of system.

Example 1: Use Inverse matrix method to solve

𝑥1 − 4𝑥2 + 𝑥3 = 6

4𝑥1 − 𝑥2 + 2𝑥3 = −1

2𝑥1 + 2𝑥2 − 3𝑥3 = −20

Solution:

4
 1 −4 1 6 𝑥1
𝐴 = [4 −1 2 ] , 𝑑 = [ −1 ] , 𝑋 = [𝑥2 ]
2 2 −3 −20 𝑥3

1/55 2/11 7/55

−1 −16/55 1/11 −2/55]
𝐴 =[
−2/11 2/11 −3/11

𝑋 = 𝐴−1 𝑑

𝑥1 1/55 2/11 7/55 6 −144/55

[𝑥2 ] = [−16/55 1/11 −2/55] [ −1 ] = [ −61/55 ]
𝑥3 −2/11 2/11 −3/11 −20 46/11

Using equality of matrices, solution of system is

144 61 −230 46
𝑥1 = − , 𝑥2 = − , 𝑥3 = =
55 55 −55 11

Example 2: Use Inverse matrix method to solve

𝑥1 + 2𝑥3 = 6

−3𝑥1 + 4𝑥2 + 6𝑥3 = 30

−𝑥1 − 2𝑥2 + 3𝑥3 = 8

Solution:

1 0 2 6 𝑥1
𝑥
𝐴 = [−3 4 6] , 𝑑 = [30] , 𝑋 = [ 2 ]
−1 −2 3 8 𝑥3

6/11 −1/11 −2/11

−1
𝐴 = [3/44 5/44 −3/11]
5/22 1/22 1/11

𝑥1 6/11 −1/11 −2/11 6 −10/11

𝑥
[ 2 ] = [3/44 5/44 −3/11] [30] = [ 18/11 ]
𝑥3 5/22 1/22 1/11 8 38/11

The solution is

5
 10 18 38
𝑥1 = − 11 , 𝑥2 = 11 , 𝑥3 = 11.

Result: If in nonhomogeneous system AX=d, |𝐴| = 0, there are two possibilities for
solution

Either infinite many solutions exist for this system or system may be inconsistent (no solution
exists)

Example 3: Solve

3𝑥1 − 𝑥2 + 𝑥3 = 4

−𝑥1 + 7𝑥2 − 2𝑥3 = 1

2𝑥1 + 6𝑥2 − 𝑥3 = 5

Solution:

3 −1 1 4
𝐴 = [−1 7 −2] , 𝑑 = [1]
2 6 −1 5

|𝐴| = 0

We cannot apply Cramer’s rule and inverse matrix method.

Example 4: Solve

𝑥1 − 3𝑥2 + 3𝑥3 = −1

𝑥1 − 3𝑥2 + 4𝑥3 = 5

−𝑥1 + 3𝑥2 + 5𝑥3 = 1

Solution:

1 −3 3 −1
𝐴 = [ 1 −3 4] , 𝑑 = [ 5 ]
−1 3 5 1

6
 |𝐴| = 0

We cannot apply Cramer’s rule and inverse matrix method. We can find solution by
Gaussian elimination or Gauss-Jordan Elimination.

Example 5: Given the following model

𝑌 = 𝐶 + 𝐼0 + 𝐺0 ,

𝐶 = 𝑎 + 𝑏(𝑌 − 𝑇) (𝑎 > 0, 0 < 𝑏 < 1) [T: taxes]

𝑇 = 𝑑 + 𝑡𝑌, (𝑑 > 0, 0 < 𝑡 < 1) [t: income tax rate]

where Y , C and T are endogenous variables, and 𝐼0 and 𝐺0 represent the exogenous
variables.

(a) Rewrite national-income model in the AX=d (with Y as the first variable in vector
X).
(b) Find Y, C and T by inverse matrix method

Solution: System can be rewritten as

𝑌 − 𝐶 = 𝐼0 + 𝐺0 ,

−𝑏𝑌 + 𝐶 + 𝑏𝑇 = 𝑎

−𝑡𝑌 + 𝑇 = 𝑑

1 −1 0 𝑌 𝐼0 + 𝐺0
[−𝑏 1 𝑏 ] [𝐶 ] = [ 𝑎 ]
−𝑡 0 1 𝑇 𝑑

where
1 −1 0 𝑌 𝐼0 + 𝐺0
𝐴 = [−𝑏 1 𝑏 ] , 𝑋 = [𝐶 ] , 𝐵 = [ 𝑎 ]
−𝑡 0 1 𝑇 𝑑

1 −1 0
|𝐴| = |−𝑏 1 𝑏|
−𝑡 0 1

Expanding along first row yields

7
 |𝐴| = 1 |1 𝑏
| − (−1) |
−𝑏 𝑏
| + 0 = 1 − 𝑏 + 𝑏𝑡
0 1 −𝑡 1

This national-income model will have unique solution only if |𝐴| = 1 − 𝑏 + 𝑏𝑡 ≠ 0.

1 1 −𝑏
𝑎𝑑𝑗( 𝐴) = [𝑏(1 − 𝑡) 1 −𝑏 ]
𝑡 𝑡 1−𝑏

1 1 1 −𝑏
−1
𝐴 = [𝑏(1 − 𝑡) 1 −𝑏 ]
1 − 𝑏 + 𝑏𝑡
𝑡 𝑡 1−𝑏

𝑌 1 1 1 −𝑏 𝐼0 + 𝐺0
[𝐶 ] = [𝑏(1 − 𝑡) 1 −𝑏 ] [ 𝑎 ]
1 − 𝑏 + 𝑏𝑡
𝑇 𝑡 𝑡 1−𝑏 𝑑

𝐼0 + 𝐺0 + 𝑎 − 𝑏𝑑
1 − 𝑏 + 𝑏𝑡
𝑌 𝑏(1 − 𝑡)(𝐼0 + 𝐺0 ) + 𝑎 − 𝑏𝑑
[𝐶 ] =
𝑇 1 − 𝑏 + 𝑏𝑡
𝑡(𝑎 + 𝐼0 + 𝐺0 ) + 𝑑(1 − 𝑏)
[ 1 − 𝑏 + 𝑏𝑡 ]

Using equality of matrices, we have

𝐼0 + 𝐺0 − 𝑏𝑑 + 𝑎
𝑌=
1 − 𝑏 + 𝑏𝑡

𝑎 − 𝑏𝑑 + 𝑏(1 − 𝑡)(𝐼0 + 𝐺0 )
𝐶=
1 − 𝑏 + 𝑏𝑡

𝑑(1 − 𝑏) + 𝑡(𝑎 + 𝐼0 + 𝐺0 )
𝑇=
1 − 𝑏 + 𝑏𝑡