UNIT 10 LINEAR ALGEBRA

Structure
Objectives
Introduction
Determinants
10.2.1 Definition and Concepts
10.2.2 Minors and Cofactors
10.2.3 Properties of Determinants
10.2.4 Product of Two Determinants
10.2.5 Adjoint and Reciprocal Determinants
10.2.6 Symmetric Determinants
10.2.7 Skew and Skew-symmetric Determinants
10.2.8 Solution of Simultaneous Equation by Cramer's Rule
10.2.9 Consistency of Equations
Matrix
10.3.1 Concept of Matrix
10.3.2 TypesofMatrix
10.3.3 Matrix Operations
10.3.3.1 Equality of Matrices
10.3.3.2 Addition and Subtraction of Matrices
10.3.3.3 Matrix Multiplication
10.3.3.4 Adjoint and Reciprocal Matrices
10.3.3.5Trace of a Matrix
10.3.3.6 Sub Matnces and Minors
10.3.3.7 Transpose of Sum or Difference of Matrices
10.3.3.8Transpose of a Product of Matrices
10.3.4 Matrix Inversion
10.3.5 Some Other Types of Matrices
10.3.5.1Orthogonal Matrix
10.3.5.2 Symmetric and Skew-symmetric Matrices
10.3.5.3 Idernpotent Matrix
10.3.6 Partitioned Matrices
10.3.7 Rank of a Matrix
10.3.8 Linear Dependence (Independence) and Rank of a Matrix
Let Us Sum Up
Key Words
Some Useful Books
Answer or Hints to Check Your Progress
Exercises

10.0 OBJECTIVES
After studying this unit you will be able to:
explain the concepts and presentation of equations in the form of
determinants and matrices;
0' evaluate their properties; and
know the techniques of solution offered by determinants and matrix
operations.
Linear Algebra and Economics 0 . 1 INTRODUCTION
Application
Analysis of economic relationships often requires inclusion of a large number
of interrelated variables into the model. Consequently, finding solution to
must of problems under investigation becomes cumbersome. Linear algebra
helps in various ways providing means for their concise presentation and
solution mechanism. In the present unit we discuss the tools of solving
simultaneous equations with the aid of matrices and determinants. The
following discussion starts with determinants, their main features and solution
techniques. The concept and operational methods of matrices are taken up
subsequently.

10.2 DETERMINANTS
10.2.1 Definition and Concepts
+
If the linear equation a , x b, = 0 and a,x + b, = 0 have the same solution,

then -=- . -=-,

a, bl , ~.e.,
a, a2 i.e., a,b,-a2b,=0. Theexpression (a,b, -a,b,)
a2 "2 bl b2 *

is called a determinant and is denoted by the symbol

Note that a , , a,, b , ,b, are called the 'elements of the determinant'. The
elements in the horizontal direction form 'm' and those in the vertical

direction form 'columns'. The determinant 1 :: has two rows and two

columns. So it is called a 'determinant of the second order' and it has 2 ! = 2

terms in its expansion of which one is positive and another is negative. The
'diagonal term' or the 'leading term' of the determinant is (a, ,b2) whose sign
is 'positive'.
Again, if the linear equations a , x + b,y + c, = 0 , a,x + b,y + c, = 0 and
+
a 3 x + b3y c 3 = 0 have the same solutions, then we have from the last two
equations by cross multiplication.

or, x= b2c3 - b3c2 c i a 3 -c3a2

= a 2 b 3- a3b2

where (a2b3- a 3 b 2 )# 0
These values of x and y must satisfy the first equation. Hence
a,(b2c3-b3c2)+ b,(c2a3-c3a2)
+ c,(a2b3-a,b2)=
The expression a,b2c3- a,b3c2+ a3b,c2- a,b,c, + a2b3c,- a 3 b 2 c , is
a, bl c,
denoted by the symbol a, b, c, and has three rows and three columns.
a3 b3 c3
Therefore, it is called a 'determinant of the third order' and it has 3! =6 terms
r
of which three terms are positive and rest are negative. These tenns may be Linear Algebra
obtained by writing a, b, c in the natural order and arranging the suffixes 1,2,
3 in all possible orders. The 'diagonal term' or the 'leading term' of the
determinant is a,, bZ,c3 whose sign is positive. So far a 'determinant of nh
order' is concerned there will be n! terms in its expansion and out of them half
of the terms would be positive and rest would be negative.
I Moreover, a third order determinant can be written as:

au -
au
A determinant of the form
ax ay = au av - --
-- a u av
, where u and v are
av
- av - ax ay ay ax
ax ay
two given functions of x and y, is termed as 'Jacobean' of u and v with respect
a(u, v)
to x and y and is often written as -
I a ( x , Y)

I
Similarly, a(u7v'W, is a 3rdorder Iacobean of u, v and w as functions of x, y
i a x , y, Z )
and z.
10.2.2 Minors and Cofactors
In the last part, i t is to be noted that a, is multiplied by a lower order

determinant (here the order is 2). is obtained by deleting the column

and row containing a,. The determinant 1 1

b2
b3
c2
c3 is termed as 'minor' of a, in

the original matrix. Same procedure will be applied for other terms also.
Thus, the minor of b2 will be obtairied by deleting the 2ndrow and 2ndcolumn.

So, the minor of any element in a nthorder determinant is thus a (n-l)thorder

determinant where n= l , 2 , 3 , .. . any finite positive number.
The 'cofactor' of any element in a determinant is its coefficient in the
expansion of the determinant. It is therefore equal to the corresponding ininor
with a proper sign.
. . The sign of the cofactor of an element in the ithrow and j~
column is
Now, a, lies in the lStrow and lStcolumn, hence its cofactor is
Linear Algebra and Economics Similarly bs lies in 3" row and Tdcolumn,
Application

: 1 :I*
I

hence its cofactor is (-11~'~

a, bl c1 I
Thus, a, b, c, = alAl + blBl + clC1= alAl + a2A2 + a3A3 (Do yourself)
a3 "3 c3
where Ai, Bi, Ci (i=l, 2,3) are respective cofactors of ai, biandci
Example 1: Find the cofactors and the value of the following determinant

Solution: Here, a1 = 3, bl = 4, cl = 7, a2 = 2, b2 = 1, c2 = 3, a3 = 7, b3 = 2, c3 =
1

Here the cofactors are Al = (-1)"'

and so on.
So, expanding the determinant by lS'row, we have its value

Similarly, we can expand the determinant by any row or column. If we

expand it by the 1'' column, then, value of the determinant will be
 10.2.3 Properties of Determinants Linear Algebra

,- Property I + A determinant is unchanged in value, if all the rows of the

/
determinant are changed into columns or vice versa, i.e.,
? '11 '12 I '11 '21 an1
a21 a22 ... a h - a12 a22 ... a,
........................
a,, a,, ... a, a,, a, ... a,
1 3 0 1 1 2 1
Example 2: 2 4 0 = 3 4 0 [following property (I)]
1 0 3 0 0 3
Property II+ The value of a determinant is unaltered numerically but
changed in sign if two rows (or two columns) are
interchanged.
1 3 0 3 1 0
Example 3: 2 4 0 = - 4 2 0 [by interchanging column (1) and
1 0 3 0 1 3
(211
4 2 0
= 3 1 0 [by interchanging row (1) and (2)]
0 1 3
I
Property III+ A determinant has zero value if the elements of one row (or
of one column) are equal or proportional to the corresponding
elements of a second row (or of a second column).
1 2 3
Example 4: 2 3 1 = O [since row (1) and (3) are identical]
1 2 3
1 2 3 1 1 3
Example 5: 2 4 1 = 2 2 2 1 = O [since column (1) and (2) are
1 2 4 1 1 4
identical]
Property IV+ If all the elements of any one row (or column) are multiplied
3
(or divided) by K, then value of the determinant is also
multiplied (or divided) by K.
K'II "12 KaIn '11 '12 a~n
a21 a22 ah = K '21 '22 '2,
Example 6:
... ... ... .........
an, an, 'nn an1 an, arm

It can also be numerically shown by considering a 3rdorder determinant

2 1 0 1 6 1 5 8
3 7 1 = 2 3 7 1 [since all the elements in the first row have a
5 4 1 1 5 4 1 1
Linear Algebra and Economics COmmOn factor 21
Application
Property V -+ If all the elements of any row (or column) of a determinant
can be expressed as the sum of (or difference between) two
numbers, then the determinant can be expressed as the sum of ,
(or difference between) two determinants. T

a, +a1 '31 + P I c, +61

Example 7: a, b, C2

a3 b3 c3
Numerically it can be shown as follow

8+2 2+5 5+4 8 2 5 2 4 5

4 9 1 = A(say) = 4 9 1 + 4 9
1 =A,+A, (say)
5 11 7 5 1 1 7 5 1 1 7

and A=10(63-11)-7(28-5)+9(44-45) =520-161-9 =350

:. A = A, + A, Hence, the proposition is proved.
Property VI + The value of a determinant remains unchanged by adding (or
subtracting) k times the' elements of any row (or column) to
(or from) the corresponding elements of any other row (or
column), where k is any given number.
a,+kb, b, c, a, b, c, kb, b, c,
Example 8: a,+ kb, b, c, = a, b, c, + kb, b, c,
a,+ kb, b, c, a, b, c, kb, b, c3

[since the value of the second determinant is zero as its 1'' and 2ndcolumns are
identical]
It can also be numerically shown as follow:
 Linear Algebra

:. A = A, = A, ; Hence, the proposition is proved.

t
10.2.4 Product of Two Determinants
The product of two determinants of same order is another determinant of the
same order.

4 5 9 2 1 5
Numerically, 1 2 3 x 3 2 1
3 0 1 1 6 4
4x2+5x3+9xl 4xl+5x2+9x6 4x5+5x1+9x4 32 68 61
= lx2+2x3+3xl lx1+2x2+3x6 lx5+2x1+3x4 = 1 1 23 19
3~2+Ox3+2x1 3xl+Ox2+2x6 3~5+Ox1+2x4 8 15 23
Check Your Progress 1
2 1 3
1) Find the value of the determinant 4 5 6
7 8 9
Linear Algebra and Economics 2 0 -1
Application
2) Your are given a determinant 1 1 7 . Work on it to verify important
3 3 9
properties.

3) Evaluate a1 b 11

......................................................................................
b+c a a
4) Find the value of b c+a b
c c a+b

10.2.5 Adjoint and Reciprocal Determinants

The 'adjoint' or 'ndjugate' of a determinant A is the determinantA ' whose
elements are the cofactors of the corresponding elements of A

a , bl c, Al Bl Cl
Let A = a, b, c, then, A ' = A, B, C,
a3 "3 C3 A, B3 C3
Where A,, BI, Cl, ...are the respective cofactors of al, bl, cl,. ...of A.
Note: Here A = A when A # 0.

0 1 2
Example 11: Let A = 2 0 1
1 2 0
Here, a, =0, b, = l,c, = 2
a, = 2 , b 2=O,c, =1
a3 =3,b3 =2,c3= O
 Linear Algebra

Sirnilarly,C, = 4 , A , = 4 , B 2 =-2,C2 =1,

A, = l , B , = 4 , C 3 =-2
-2 1 4
.: Adjoint of Ais A'= 4 -2 1
1 4 -2

The 'reciprocal' or 'inverse' of a determinant A(#O) is the determinant,

which is formed by dividing every element of the adjoint of A by A . Thus,
the reciprocal of A is
435

Example 12: Following the last example (i.e., Example 1I), we get,
A=O(O-2)+1(1-0)+2(4-0) = 9 + 0

l9
:.The reciprocal of A is 419 -219
9 9
119 I
10.2.6 Symmetric Determinants

If a determinant is written in the form a,, 1 a,, 1

aZ3

in which the suffix of any element indicates its position (e.g., the suffix 12 of
the element a12 indicates that the element is in the IS' row and in the 2nd
column) and if aij = aj, for all i j = 1, 2, 3; then the determinant is said to be
symmetric.

a h g
Thus h b f is a symmetric determinant.
g f c
Properties: <a>?-'hesquare of any determinant is a symmetric determinant.
<b>The adjoint of symmetric determinant is symmetric.
Lini!ar Algebra and Economics
Application

Here A is a symmetric determinant

14 25 34
= 25 45 61 , which is also a symmetric determinant.
34 61 83

l3
Now adj. A= 1 -2 2~
1 , which is also a symmetric

determinant
10.2.7 Skew and Skew-Symmetric Determinants

The determinant a,, a,,

is said to be 'skew', if aij=-aji V ij=1,2,3 and i #j .

-b
a -g
Thus, h b - f is a skew determinant.
g f c
And if aij=-aji V ij=1,2,3 and i #j with aij = 0 'di = j,
then the determinant is said to be skew-symmetric.
o a b
Thus, - a o c is a skew-symmetric determinant.
-b -c o
Properties: <a>Every skew-symmetric determinant of third order (more
generally of odd order) is zero.
<b>Every skew-symmetric determinant of second order (more
generally of even order) is a perfect square.
 0 5 2 0 -5 -2 Linear Algebra

Example 14: LetA = -5 0 -3 = 5 0 3

-2 3 0 2 -3 0
[taking (-1) common from each column]
= -A
Hence, 2A = 0 3 A=0

Again, let A, =
I-O2 :I
:.A, = 0 - (-4) = 4 = 2, which is a perfect square.
10.2.8 Solution of Simultaneous Equations by Cramer's Rule
Gabriel Cramer obtained a simple method of solution of linear simultaneous
equations by using the properties of determinants.
In particular, let the system of equations be

xa, b, C, a , x + b , y + c , z b, c,
Then xA = xa, b, c, = a,x + b,y + c,z b, c, [C,'=C,+YC,+
zc3]
xa, b, c, a3x +b3y +c3z b3 c3

kl b, c,
= k, b, c, [from(l)]=A, (say)
k3 b3 C3

A whereA, and A, are the determinants obtained

Similarly, y =-A2 and z=--1-
A A
from A substituting the columns (kl, k2, k3) for the second and third column
respectively.
Example 15: Suppose that the numbers x l and x:! satisfy the equations
XI- 2x2 = 3 and 3 x 1 + 5x2 = 20.
Find xl and xz by using Cramer's rule.
Solution: X I - 2x2 = 3,3x1 + 5x2 = 20
It may be written as AX = B

where, A=( 3
1 -2
)
5 ,AX = H X=(::) and B=(lo)
Linear Algebra and Economics Cramer's rule tells us that
Application

13 201 l(20)-3(3) -
- -11
and x, =
11 - 2 / = 11 11 = I

Thus, if we have the value of the original determinant along with the values
obtained from substituting the column of constants (i.e., ki) in different places,
then by Cramer's rule we can have the value of the corresponding variable.
10.2.9 Consistency of Equations
When a system of linear equations has one or more solutions, the equations
are said to be 'consistent' or 'compatible'; otherwise they are said to be
'inconsistent' or 'incompatible'.
For the system of equations six + biy + c,z = ki (i=1,2,3)
we have, by Cramer's Rule, the unique solution
A A2 A3
x ,=' y =-, z =- provided A # 0 and A-,(i=l, 2, 3) is the determinant
A A A
obtained by replacing the ith column in the coefficient determinant by the
column consisting of elements kl, k2, k3.
Thus, the above equations are consistent, if A # 0 .
If A = 0 and at least one of the determinants A,, A , , A, be non-zero, then
these equations do not posses any solution and are therefore inconsistent.
On the other hand, if A = 0 and all of A,, A,, A, are also zero. the system
admits of an infinite number of solutions.
Note: If kl=k2=k3=0, then the equations reduce to the set of homogeneous
linear equations six + biy + ciz = ki (i=l, 2,3)
0
By Cramer's rule, in this case, we have x=y=z=-
A
If A # 0 , then the equations have no other but obvious solution x = y = z = 0,
which is called a 'trivial solution'.
On the other hand, such solutions in which x, y and z are not all zero are
called a 'non-trivial solution'.
Check Your Progress 2
1) Determinelhe signs to be attached to relevant minors if you are given
the co-factors of a determinant as IC131,IC23), JC331,IC411 and IC34(
2) Solve by Cramer's rule x + 24 + 32 = 6 Linear Algebra
2x+4y+z=7
3x+24+9z= 14

10.3 MATRIX
10.3.1 Concept of Matrix
In real life, we do not have to deal with a single magnitude but with a set of
magnitudes. In analysis we treat such a set of magnitudes as a single entity or
a single object of thought, abstracted by numbers. Let us consider a problem
where we treat a set as single entity and a number of such sets as a single
system. We will then examine how the fundamental operations of arithmetic
can be suitably defined to deal with such a system.
Suppose there are 3 inciividuals A, B, C in a family and that
A has a set of 3 pants, 3 shirts and 1 tie.
B has a set of 5 pants, 5 shirts and 2 ties.
C has a set of 6 pants, 8 shirts and no tie.
We can arrange the data in the following convenient system:
Individuals Pants Shirts Ties
.1 (f) (S) (TI

A 3 3 1" Row
3 2nd Row
C 3 3 3' Row
.1 L L
1" 2 nd 3rd

Col. 0 Col.
This system comprises 3 rows and 3 columns (note that, by convention, rows
are written first and coiumns afterwards). We have written belongings of A in
the first row, belongings of B in the second row and the belongings of C in the
third row. Clearly, therefore, IS' column gives us the total member of pants
that A, B and C together have while second and third columns enumerate
respectively the number of shirts and ties that three individuals together have.
Number written in such a particular form of rows and columns and enclosed
by [ ] 3 square brackets or ( )+large parentheses are called a 'matrix'.
A 'matrix' is defined as a.rectangular array of elements arranged in rows and
columns and 'matrix algebra' basically provides methods for dealing with sets
of simultaneous relationships involving many variables.
Linear Algebra and Economics
Application
In general form,
=
'In2
a22
.. ::"
...'ma,
1
This is a matrix of m rows and n columns. It is denoted by (aij) and is read as
m by n. The first subscript refers to the number of row and the second to the
number of column to which the element belongs. Thus, a13 belongs to the lS'
row and 3rd column; in general, aij refers to the element of jth row and it"
column. If the matrix has total number of elements = mn, arranged in m rows
and n column, it is said to be of the order of m by n, which is often written as
m x n matrix.
10.3.2 Types of Matrix
Rectangular Matrix
A matrix of the order m by n is also called as 'rectangular matrix'. And if
m=n, then the matrix is said to be a 'square matrix' of order 'n'.
Column Matrix

If the matrix consists of only one column as

1: 1
La31
, then it is called a 'column

matrix' or 'column vector'.

Null Matrix
If all the elements of a matrix be zero, then it is called a 'null-matrix' or 'zero
matrix'. The null matrix of order m x n is denoted by Om,.

If the elements of a matrix be not all zero, then it is called a 'non-zero matrix'
The negative of a matrix is [-aij] ,,, and is denoted by (-A).

Thus, if
"I c,
A=!:: b, c2]

The 'transpose of a matrix' A is the matrix obtained from A by changing its

rows into columns and columns into rows. It is denoted by A' or AT. Hence if
A be of order (m x n), then ATis of order (n x m).
F 7

Thus, the transpose of

Note that, r
[ A ~ =A
If the number of columns be equal its number of rows, then the matrix is Linear Algebra
called a 'square matrix'. There are some special forms of square matrices,
viz.,
Diagonal Matrix
Jf all off-diagonal elements of a square matrix are zero, that matrix is called
diagonal matrix. Thus

(a,,, ... a,,, I

ai, = 0 for i # j and aij# 0 for at least one i = j

I
For example, A = 0 2 0 is a diagonal mamx.

Scalar Matrix
If the diagonal elements of a diagonal matrix be all equal, then the matrix is
called a 'scalar matrix'.

1: :I:
e.g., 0 a 0 is a scalar matrix [a]

Identity Matrix
If the diagonal elements of a diagonal matrix are all equal to 1, it is defined as
a 'unit matrix' or 'identity matrix'. A unit matrix of order 'n' is denoted by I,
or by [I] or simply by I. Thus,

A= [I:.1: ] = (aij),.,, is an identity matrix if and only if

a,,=Ofor i # j a n d a i j = l f o r i = j

Example: I3= is an identity matrix of order 3.

Note that I = I'

10.3.3 Matrix Operations
10.3.3.1 Equality of Matrices
Two matrices A and B are said to be 'confomable for equality' if they were
of same order.
Two matrices of same order are said to be 'equal', if and only if, the
corresponding elements of the two are equal.
1 4
Example 16:
Linear Algebra and Economics If the two matrices are not of the same order, then they are 'not confomable
A~pUcation for equaliry' .
10.3.3.2 Addition and Subtraction of Matrices
If the matrices A and B are of the same order (m x n), then they can be added
or subtracted.
] ]
If A= [aij and B= [bij], then A+B = [aij+ bij] and A-B = [aij-bij
In other words, corresponding elements are added or subtracted. The new
matrix will be of the same order (m x n).
1 3 5
Example:

Some Properties:
i) A + B = B + A (so, matrix addition is commutative)
ii) (A + B) + C = A +(B +C) where C is the third matrix (so, matrix addition
is associative).
Note: [A + B] = + B~
iii) A + 0 = 0 + A = A, thus zero matrix plays the same role in matrix algebra
as zero in ordinary algebra.
Some Results:
a) A+(-A)=(-A)+A=O
b) A + A = 2 A
c) k ( A + B ) = k A +kB
d) (kl + k2) A = klA + k2A
10.3.3.3 Matrix Multiplication
If the number of columns of a matrix A be equal to the number of rows of
another matrix B, then the matrices A and B are said to be 'conformable for
the product' AB and the product AB is said to be defined.

then AB= [c, ,I =x

where cik
n

j=l
aij bjk

Thus,

Some Properties:
i) In general AB # BA [so, matrix multiplication is not communicative in
general.]
ii) For conformable matrices (AB) C=A (BC) [so, matrix multiplication is
associative.]
iii) (B+C) D= BD+CD [so, matrix multiplication is distributive with respect to
addition]
iv) (AB) = BT .AT " Linear Algebra

10.3.3.4 Adjoint and Reciprocal Matrices

If A = [a,]. .. be a square matrix, then the transpose of the matrix [A,]] nxn
whose elements are the cofactors of the corresponding element in is called IA/
[A,]. .. . If A = [a,], ..
the 'adjoint' or 'adjugate' matrix of A and is denoted by adj.A It is equal to
be a non-singular square matrix, the matrix
adj.A
is called the 'reciprocal matrix' of A.
nxna

Thus, if A = a,,

A2 I

A,, I
[where Aijis the cofactor of ai, in ( A ]
A23 4 3
- -
-
All - A21 -
A31

IAI IAl IAl

and the reciprocal of A is A,, A,, A 32
-
IAl IAl IAl
A,, A,, A,,
-l~l IAl
10.3.3.5 Trace of a Matrix
The trace of a square matrix is the sum of its diagonal elements.

Thus, in the matrix the trace (tr) is = 1 + 0 - 1=0

In general, trace tr (A) = a,, + a,, +....+ a, = n aii

i=l

Properties: i) tr (A+B) = tr (A) + tr (B)

ii) tr(A')=tr(A)
iii) tr (AB) = tr (BA) and
iv) tr (ABC) = tr (BCA) = tr (CAB).
10.3.3.6 Sub Matrices and Minors
Any matrix obtained by omitting some rows and columns of a given m x n
matrix A is called 'sub matrix' of A.

Thus, [: :]is a sub matrix of

Linear Algebra and Economics The determinant of the square matrix of order 'r', obtained from a given m x n
Application matrix A by omitting (m - r) rows and (n - r) columns, is called a 'minor' of
A of order r' .
I

10.3.3.7 Transpose of Sum or Difference of Matrices

The transpose of a sum or difference of matrices is equal to the sum or
difference of the transpose of the matrices. Thus,
(A,,,* B,, * C,,)' ',, + B',,
=A * C',m,
Example:
2 -1
A = [ 0 -1 ] 6 8 0 -1
B=[ 10 -12] C = [ 0 0 ]then

[A + B + CT = [-10
-4 -10
11] [
-10
-4
= -10 11
]
or alternatively

10.3.3.8 Transpose of a Product of Matrices

The transpose of a product of matrices is equal to the product of the transpose
of the matrices in inverse sequence; thus

Example: 3 -1
A = r -4 '1 3 5 -7
B = [ 0 -1 8 ] C=[-:]then

10.3.4 Matrix Inversion

I I
A square rnaoix An, ,with A f O is called a non-singular matrix. An n is
1 I
called singular matrix if A = 0.

1
The inverse of a non-singular matrix A is given by A-' = -.Adj.A
IAl
1 1
Example 17: A = :. l~I=l(l-6)-I(-2-9)+1(4+3)
 Linear Algebra

So A is a non-singular matrix, its inverse exists.

Now, matrix of cofactors of A=

Some Properties:
i) The inverse of a matrix, if exists, is unique.
ii) The necessary and sufficient condition that the inverse of a square
matrix A exist, is that A is non-singular.
iii) [ABI-' =B". A-' (this is known as 'law of reversal of inverse')
iv) (A-')=A
V) (A-I) = ( ~ ~ 1 - l
and (A")-' T. AT = I = AT (A-I) (operations of reciprocating and
transposing are commutative)
Check Your Progress 3

1) Obtain the inverse of A = 1 3 ( 5 1

3) Given a matrix, [ 7 -1
1 1
13 -3
0
4 1 , what would you conclude about its
-4
rank?
Linear Algebra and Economics 10.3.5 Some Other Types of Matrices
Application
10.3.5.1 Orthogonal Matrix
A square matrix A is said to be orthogonal, if its transpose produces a unit
matrix when multiplied by A, i.e., A ~A. = 1.
Some Properties:
i) Orthogonal matrices are non-singular,
i.e., if A=.A = I then A # 1 01
ii) Unit mamces are orthogonal, i.e.,
IT. I = I . I = I
iii) The value of the determinant of an orthogonal mamx is either 1 or (-1).
If A is an orthogonal matrix, then A =I

iv) The product of two orthogonal matrices is orthogonal.

v) The transpose of an orthogonal matrix is orthogonal.
vi) The inverse of an orthogonal matrix is orthogonal.
10.3.5.2 Symmetric and Skew-Symmetric Matrices
A s uare matrix A is said to be 'symmetric' if its transpose be equal to it, i.e.,
9r
if A =A. Hence, the matrix [aj] is symmetric, if aij = aji for all i and j. For
example,

[i !]is asymmetricmatrix.

A square matrix A is said to be 'skew-symmetric', if its transpose is equal to

its negative. i.e., if = -A. Hence, the matrix [aij] is skew-symmetric, if a,, =

,
-aji for all i and j. The diagonal elements of skew - symmetric matrix are
obvious1;r zero. For example,

[-::]
-b -C 0
is a skew - symme,c matrix.

Some Properties:
i) The product of a mamx and its transpose is a symmetric matrix
Now, [A A ~ ]= [ A ~T.] [A] = A. [A]
Hence, A . A is
~ symmetric.
ii) The product of two symmetric matrices of the same order is
symmetric if the product by commutative.
If A and B be two symmetric matrices of the same order and AB = BA,
then [ A B ] =
~ B ~ . =AB.A
~ = AB.
Hence, AB is symmetric.
iii) If A is a square matrix, then (A + AT) is symmetric and (A - AT)is Linear Algebra
skew-symmetric.
(A + = +( A ~=)A~~ + A
Thus, (A + AT) is symmetric.
Similarly, (A - ATIT= AT - (ATIT= - A = -(A - AT)
Hence, (A - A ~ is) skew-symmetric.
iv) Every square matrix can be expressed as a sum of a symmetric
matrix and a skew symmetric matrix uniquely.
Let A be a square mamx, then,

10.3.5.3 Idempotent Matrix

A symmetric matrix that reproduces itself when multiplied by itself is termed
as an idempotent matrix.
That is, A will be termed as idempotent matrix if AA = A.
10.3.6 Partitioned Matrices
It is often convenient to partition a mamx into sub matrices. These sub
matrices can then be treated as scalars in performing operations on the original
matrix. Partitioning of a matrix is indicated by horizontal and vertical lines
between rows and columns.
The mxn matrix A may be partitioned as follows A = (A1 i A2)
If they are partitioned conformably, partitioned matrices can be added,
subtracted or multiplied. If an mxn matrix A is partitioned A = (A1 i A2),
where A1 is mxnl, A2 is mxn2 and nl+n2=n, and mxn matrix B is partitioned B
= (B1 B2), where B1 is mxnl, B2 is mxn2 and nl+n2=n, then
A * B = [A,*B, i A,*B,]

I: [
Similarly, if A =

m1+m2=m, and B =[::I,

-.. , where A is mxn, A] is mlxn, A2 is m2xn and

where B is mxn, B is mlxn, B2 is m2xn and

ml+m2=mthen [ ]
Al* B,

A,
-
* B,
Partitioning is frequently conceptually convenient when matrices are to be
added or subtracted, but the computational advantages of matrix partitioning
are associated primarily with multiplication and other more complex
operations. Partitioned matrices must be partitioned conformable for
multiplication. If matrix A of order mxn is partitioned A = (Al i A2),where
Al is mxnl, A2 is mxn2 and nl+n2=n and nxp matrix B is partitioned
Linear Algebra and Economics
Application
B =[%:], where B is nlxp and B; is n ~ x pthen

AB = [ A : A21[~:]= A2B2

Another example: if A =

A21 is m2xn1,
[
A21
All
.
i
;.
A22
. .I
A12
where A is m 1 x n A12is mlxn2,

is m2xn2 and rnl+mz=m, nl+n2=n and if

B= [?: . q:], where Bll is nlxpl, B12 is nlxp2, B21 is n2xp1, B22 is
B2l B22
n2xp2 and nl+nz=n, p1+p2=p,then A and B are partitioned conformably for
All i 4 2 Bll i Bl2
multiplication and AB =

Check Your Progress 4

1) List the Properties of orthogonal mamx.

......................................................................................
2) Define an idempotent mamx.

A= 1; ::!
(1

0
3

2
1)

-3
(3 -2 1 0
and B = 15 -1 4 -3 21
3 -2 0 1 -1
-1)

then show AB =
 Linear Algebra

10.3.7 Rank of a Matrix

A number 'r' is said to be the rank of a matrix A, if
i) there is at least one (r x r) sub-matrix of A whose determinant is not equal
to zero; and
ii) the determinant of every (r+l) - rowed square sub-matrix of A is
zero.
In other words, rank of a matrix is equal to the order of the highest order non-
singular square matrix contained in A. Moreover, rank of a matrix can also be
defined as the maximum number of rows (or, columns) in a matrix.

For example, the rank of matrix [-52 -23] is 2; because this second order

matrix corresponds to the determinant [-52

L
-2A.
J
which is not zero, hence

this given matrix contains a non-singular matrix of order 2.

Whereas, the rank of the matrix :/ is 1; the highest order determinant

1 is zero, next to this would be of the 1" order, which is non-zero,

Note:
[::I:
Again, the rank of 0 0 0 is zero.

i) The rank of a matrix cannot exceed the number of its rows or columns -
whichever is less.
ii) The rank of a column or row matrix is at most 1, unless the matrix is a null
matrix, where the rank is always zero.
iii) Rank of the transpose of a matrix is the same as the rank of that matrix.

Example 18: Determine the rank of the matrix . Here,

L

picking any sub-matrix of order 2, it can easily be shown that

value of thk determinant is non-zero, hence rank = 2.
There is another method of determining rank
cI14c-
l 2c4
2-1 3 1 51-[00 0 1 5 1, by, C2 ' + C Z - C 4
3 2 0 2 1 - 1 4 - 6 2 1 c,'-+c~- 3c4
Linear Algebra and Economics
Application -[-:-::-:I, by, c5'-+c5 -5c.

c3"+9c3'-6c5'
0 0 0 1 0
b y , R 2 ' 3 R 2-2R1
- [ o o o o ~ ]

So, rank of the matrix is 2

Check Your Progress 5
1) Show that rank of A=2 and rank of B=2, where

but rank of (A+B)=3.

Show further that, rank of (AB) = 1, while rank of (BA) = 2.

......................................................................................
2) How would you determine linear dependence of matrix?

......................................................................................
3) Define the rank of matrix in terms of linear in dependence.

10.3.8 Linear Dependence (Independence) and Rank of a

Matrix
The vector of rows (columns) of a matrix are said to be linearly dependent if
and only if some linear combination of them is a null vector or null row
(column). The rows (columns), which are not linearly dependent, are said to
linearly independent.
Consider the matrix:
A linear combination of the rows 1, 2 and 3 are obtained by multiplying the Linear Algebra
I"', 2ndand 31d rows by any constants kl,k2 and k3 and adding them, where at
least one constant is not zero. Thus, the new row will be

= [CI c2 ~ 3 (say)
1

The rows are linearly dependent if all the three elements cl, c2 and c3 are equal
to zero for some values of kl,k2, k3. Of course, kl, k2, k3 should all not be
equal to zero, at least one of these must be non-zero.
With regard to the concept of linear dependence (independence), rank of a
matrix can also be defined as follows:
The rank of matrix A (denoted as p(A),), is the maximum number of linearly
independent rows (or columns) in A.

10.'4 LET US SUM UP

Through this unit we have learnt how by using matrix and determinants
method an equation systknl can be written in a compact way. Moreover, we
studied the way of testing the existence of a solution by evaluating matrices
and determinants and finding the solutions (if exist) of n equation system.

10.5 KEY WORDS

Cofactor: The 'cofactor' of any .element in a determinant is its coefficient in
the expansion of the determinant. It is therefore equal to the corresponding
minor with a proper sign. The sign of the cofactor of an element in the ithrow
and jth column is (-1)"'.
Cramer's Rule: In particular, let the system of equations be

And A = a, b, c, $ 0
b1

Ixa1 cl
Then x A = xa, b1 c,
b,

a,x + b, y + c,z b, c, '

a,x + b,y + c,z b, c, [cia=c, +yc, + zc3]

a , x + b 3 y + c 3 z b3 c,
kl bl c,
= k, b, c, [from(l)] = A , (say)
k3 b3 C3
Linear Algebra and Economics A where A, and A, are the determinants
Application Similarly, y = -
A2
and z =--2-
A A
obtained from A substituting the columns (kl, kl, k3) for the second
and third column respectively? Thus, by Cramer's rule variables of an
equation system can be obtained.
Determinants: If the linear equation a l x + b, = 0 and a 2 x+ b2 = 0 have the
a , - b ,, i.e., -
same solution, then -- a, =L,
a or, a,b, -a,bl =O . The expression
a2 b2 "1 b2

: 1 ,":I-
(a,b2- a 2 b , ) is called a determinant and is denoted by the symbol

Inverse of a Matrix: A square matrix an ..

with ( A ( to is called a non-
singular matrix. The inverse of a non-singular matrix A is given by

Linear Dependence: The vector of rows (columns) of a matrix are said to be

linearly dependent, if and only if, some linear combination of them is a null
vector or null row (column).
Matrix: A 'matrix' is defined as a rectangular array of elements arranged in
rows and columns and 'matrix algebra' basically provides methods for
dealing with sets of simultaneous relationships involving many variables.
Minor: The 'minor' of any element in a determinant is another determinant of
lower order that can be obtained by deleting the row and column containing
that element. So in a 3rdorder determinant minor of any element will be a 2nd
order determinant.
Rank: Rank is the maximum number of independent rows (or, columns) in a
matrix.
Trace: The trace of a square matrix is the sum of its diagonal elements.
10.6 SOME USEFUL BOOKS
Chiang Alpha C. (1984), Fundamental Methods of Mathematical Economics,
Third edition, McGraw-Hill Book Company.
Henderson J.M. & Quandt R.E, Microeconomic Theory: A Mathematical
Approach, Third edition, McGraw-Hill Book Company.
Chakravorty, J.G & P.R. Ghosh (1996), Higher Algebra including Modern
Algebra, Twelfth Edition, U.N. Dhur & Sons Pvt. Ltd. Calcutta.

10.7 ANSWER OR HINT TO CHECK YOUR

PROGRESS
Check Your Progress 1
1) -9
2) Do yourself.
 1 0 ' 0 L i a r Algebra
= a b-a
a 2 b2 - a 2
c-a
C 2 -a2 I C, =c,
C3 = c 3-C1
-C1

1
b-a c-a
=1 2 [Expanding by rows]
b 2- a c2-a

(b+c)-(b+c) a-(2c+a) a-(2b+a)

4) [ Hint :given determinant = b c+a b
c c a+b

'0 c b O c b
( R ; = R2 - R, 9
=-2 b C + U b =-2 b a 0
c c a+b c O a
LR;
= R, - R , 1)

O b c
= + 2 b 0 a = 4abc (after evaluating)
c a O
Check Your Progress 2
1) Do yourself read Section 10.2.
1 2 3 6 2 3
2) [ ~ i n t :A = 2 4 1 = - 2 O t O : . A , = 7 4 1 =-20
3 2 9 14 2 9

Check Your Progress 3

2) Follow Example 17
3) Rank is less than 3
Check Your Progress 4
1) See 10.2.5.1
2) See 10.2.5.3
Linear Algebra and Economics 3) First partition the matrices A and B then, separately determine [AIi B I ~+
Application A12B211r [A11B12 + A12J32219 [A21B11 + A22B211 and [A21B12+A22B221
Check Your Progress 5
1) Follow Example 18
2) See 10.2.7
3) See 10.2.7

10.8 EXERCISES
1) Use Cramer's rule to solve the following national income model:
Y = C + I + G And C = a + b Y
[Hint: Find Y and C, consider I, G, a, b, as constants]

Answer: Y =
I+G+a
, c=a + b ( I + G )
1-b 1-b
2) Consider the following national income determination model:
Y=C+I+G, c=a+b(Y-T), T=d+tY
Where Y (national income), C (consumption expenditure) and T (tax
collection) are endogenous variables, I (investment) and G (government
expenditure) are exogenous variables; t is the income tax rate.
Solve for the endogenous variables, using Cramer's rule.

Ans : y =
bd-(I+G+a)
, C=
-a + bd - b(1- t)(I + G)
b(l = t) - 1 b(1 -f) - 1

3) The following is an example of the Keynesian model with the money.

Behavioral equations are
C = 0.8Y, I = 102-0.2r,
M~ = 0.25Y - 2 3 , MS=lOO
Equilibrium conditions are Y = C + I and M~= MS
ii) Write down the equations of the IS and LM curves.
iii) Evaluate the equilibrium value of Y and r using Cramer's rule.

4) Solvetheequations: x + y + z = l , x + 2 y + 3 ~ = 6 ,and x + 3 y + 4 z = 6 .
5) A set of linear equations in which the right-hand side consisting of
constants has zeros, is called a set of homogeneous equations.
Showthat a x + b y + c z = O , b x + c y + a z = O , c x + c y + b z = O

a b c
with D = b c a
c a b
 has the trivial solution (x = y = z = 0) if D # 0 and has a non-trivial Linear Algebra
solution
if a 3 + b 3 + c 3 - 3 a b c = 0 .
Again, find the solutioll treating x to be constant.
2 -3 4
6) Find the inverse of the matrix and hence obtain the

solution of the equations

2x1-3x, +4x3 =-4, XI + x 3 = o , -x,+4x3 =2.
7) Consider the national income model given by
Y=C+I+G
c=a+b(Y-T) (a>0, O c b c l )
T=d+tY (diO, O<t<l)
Solve this model for the variables Y, C and T by matrix inversion.

then find a. UV; b.VU+X; c. XY

Use partitioned matrices to check the results.
9) If A is a non-zero (n x 1) column matrix, while B is a non-zero matrix,
then show that P(BA) is unity and P(AB) is n.

10) Show that rank of the transpose of the matrix 1 -3

1
2 :1 is the

same as that the original matrix.

UNIT 11 INPUT-OUTPUT ANALYSIS
Structure
11.0 Objectives
11.1 Introduction
11.2 Main Features of Input-Output Analysis
11.2.1 Assumptions
11.2.2 Input-Output Table
1 1.2.3 Technological Coefficient Matrix
11.3 Closed and Open Input-Output Models
11.4 Coefficient Matrix and Open Model
11.5 Solution to Open Model
11.5.1 Determination of Gross Output
11.5.2 The Hawkins-Simon Conditions
11.5.3 Consumption Possibility Locus
11.5.4 Determination of Equilibrium Prices
11.6 Linear Programming Formulation
11.7 Coefficient Matrix and Closed Model
11.8 Dynamic Input-Output Model
11.9 Limitations of Input-Output Analysis
11.10 LetUsSumUp
1'1.11 Key Words
11.12 Some Useful Books
11.13 Answer or Hints to Check Your Progress
11.14 Exercises

11. OBJECTIVES
After studying this unit, you will be able to answer the following:
with what proportions one sector of the economy are related to other
sectors;
how solution is attained in a framework of several variables production
problems related to input-ou tput; .
how the system of equations input-output can be represented graphically
by the linear programming; and
how the system of input-output can be solved in a dynamic framework.

11. INTRODUCTION
The concept of Input-Output (1-0) analysis originated by the eighteenth
century French economists Quesnay in the name of 'Tabelau Economique'. In
its modern form, the 1-0 analysis has been developed by the American
economist Wassily W. Leontief in his famous work 'Structure of the
American Economy' in the year 1951.
1-0 analysis is a method of analysing how an industry undertakes production
by using the output of other industries in the economy and how the output of
the given industries used up in other industries or sectors. Since various
industries are interdependent, i.e., the output of one industry is an input for the