Chapter 4

The input - output model (Leontief model)

4.1. Introduction Matrices have proved their importance in quantitative analysis of decision
making in various fields like marketing, finance, economics etc.
Most of the quantitative methods such as the input – output model have matrix algebra as their
underlying theoretical base.
The input – output model is built by establishing a system of linear equation which represents the
problem to be solved. In this particular chapter, we will see the application of matrices in Input –
output model.
Objectives
After completion of this chapter, you should be able to
 Explain input – output model
 Differentiate the technology matrix from the Leontief matrix
 Apply matrix algebra on the Leontief model.
 The input output model
The input – output model describes the economy as a flow of goods and services among the
different sectors of the economy plus final consumption. It mainly deals with a particle of
Question “What level of output should be produced in each sectors of an economy in order that it
will just be sufficient to satisfy the total demand for that product”

In this model, there is sectoral interdependence, that is the final output of one sector is the input
for the other sector or inputs are intermediate in nature.

Thus, the essence of input – output analysis is that given certain technological coefficients and
final demands, each exogenous sector would find its output unequally determined as a linear
combination of multi – sector demand. Let us suppose that an economy consists of n producing
sectors. In order to avoid bottlenecks in the economy, the total output of each producing sector
must satisfy the total demand for its product which in fact would arise due to:

1|Page
 i- Its product is being used as an intermediate input elsewhere in the industrial structure of
production.
ii- Its product is used for household consumption, capital formation Government
consumption on for exports
For example, total output of agricultural sector may be used ( or demanded)
a) as an input is food or other manufacturing sector (grain for producing bread or cotton for
producing cloth, and
b) as a final consumption by government or household (vegetables or grain) and /or as an
export demand.
Let us assume that an economy consists of 4 producing sectors only and that the production of
each sector is being used as an input in all sectors and is used for final consumption.

Suppose : i)  1
,  ,2 3
and  4
are total outputs of the 4 sectors.

ii) F 1
, F 2
, F 3
and F 3
are the amount of final demand consumption capital

formation and exports for outputs of these sectors.

Now we can translate the distribution of total product of 4 producing sectors in the following
way.
Input - output transaction table
Producing Total output of Input sector requirements of producing Requirements for
sector No. the sector  1
 2
 3
 4
final demands

1  1
 11
 12
 13
 14
F 1

     F 2
2 2 21 22 23 24

3     
3 31 32 33 4
F 3

4
 4
 41
 42
 43
 44 F 4

Primary Total primary L 1 L 2 L 3 L 4

__
input input(Labour)
(labour) (L)

2|Page
 1
  11
 12

13
  14
 F 1

 2
 
21

22
 
23
 
24
 F 2

 3
 
31

32

33
 
34
 F 3

 4
 
41

42

43
 
44
 F 4

L  L 1
 L 2
 L
3
 L 4

In general terms, if there are n producing sectors

 1
 11
  12
 13

  
m
 F 1

 2

21
  22

23
 
2n
F 2

      
 n
  n1
  n2
  n3
  
nn
 F n

L  L 1
 L  L   L
2 3 n
n n
That is  i
  ij  F and L   Li
j 1
i
i 1

Where  i
 total output of the i
th
sector

 ij
= output of the i
th
sector as input used in i
th
sector

th
F i
= Final demand for the i sector.

The above identity states that all the output of a particular sector could be used either as an input
in one of the producing sectors of the economy and /or as final demand.
4.2. Assumptions
The input output analysis is based on the following basic assumptions.
1. Each industry produces only one homogeneous output. No two products are produced
jointly; but if at all there is such a case then it is assumed that they are produced in fixed
proportion.
2. Each producing sector satisfies the properties of linear homogeneous production function,
that is production of each sector is subject to constant returns to scale.
3. After stronger assumption is that each industry uses fixed input ratio for the production of
its output; in other words, input requirements per unit of output in each sector remain
fixed and constant.

3|Page
4.3.The Technological Coefficient matrix.
th
From the assumption of fixed input requirements, we see that in order to produce one unit of j
th
commodity the input used of i commodity must be at fixed amount, which we denote by a ij
;

a   

ij th
j
ij
thus if represents the total output of the commodity, the input
 i
j

a  a
th
requirements of i commodity will be equal to ij
or  ij
.
j ij j

As such we can now put the input – output transaction table in terms of technical coefficients as
follows:

Producing Total output of Input requirements of producing sectors Requirements

Sector No. the sector  1
 2
 3
 4
for final uses

1  1 a  11 1 a  12 2 a 
13 3 a  14 4
F 1

2
 2 a 21 1 a 22  2 a 23 3 a 24  4
F 2
3  a 31 a 32  a 33 a 34 
3 1 2 3 4
F 3

4
 4 a 41 a 42  1 2 a 43 3 a 44  4 F 4

Primary Total primary

L 1 1 L 2 2 L
3 3 L 4 4
__
input (L) input
The above table gives us the total output of each sector in terms of technical coefficients.
The above table can be summarized as:

 1
 a  a
11 1 12  2
 a   a  F
13 3 14 4 1

 2
 a  a 21 1 22  2
 a   a  F
23 3 24 4 2

 3
 a  a 31 1 23  3
 a   a  F
33 3 34 4 3

 4
 a  a 41 1

24 4
a   a  F
34 4 44 4 4

L = L 1 1
 L
2 2
 L 3 3
 L 4 4

4|Page
 4 4
That is ,  i
 a 
j 1
ij j
 F , i  1 , 2 , 3    , 4 and
i
L L 
i 1
i i

The above system can be put in matrix form as

   
 1 a11 a a
12 13 a 14
  1 F1 
       
 2 a21 a a 22 23 a 24   2  F 2
     
   
 3 a31 a a 32 33 a 34 
 3 F 3
   
a41 a a a  F 4
  4   4
  42 43 44 
 

     F and
L    i i

4.4. Closed and open input - output models

In the above example, besides n industries, our model contains exogenous sectors of final
demand which supplies primary input factors (labour, services etc) which are not produced by n
industries) and consumes the outputs of the n – producing industries) (not as input) such an
input – output model is known as open model. It includes exogenous sectors in terms of “final
demand Lill” along with the endogenous sectors in terms of n – producing sectors.
Input – output model which has endogenous final demand vector is called closed input – output
model.
4.5. Coefficient matrix and open model
The open - I - O model in matrix notation is given by :
    F
Where A is the input coefficient matrix
F is the final demand vector
X is the total output matrix
The input coefficient matrix represented by a  is given as
ij

a11 a      a1n 
12
 
a21 a22
     a2 n 
 
   
a an 2      ann
 n1

5|Page
Each column of the above matrix specifies the input requirement for the production of one unit
of a particular commodity. The second column, for instance, states that to produce a unit of ( Birr
worth of ) the second commodity the inputs needed are a 12
units of commodity 1, a 22
unit of

commodity 3 and a n2
units of commodity n

If no industry trades with each other ( if no industry uses its own product as an intermediate
input, then
0 a12     a1n 
 
a 21 0     a2 n 
a ij
  
   
a 
 n1 an 2    0 
Here, in view of the presence of exogenous sector (which supplies primary inputs) the sum of the
elements of each input coefficient column ( aij) must be less than L. Each column sum –

represents the partial input cost (excluding cost of primary input) incurred in producing a Birr
worth of a commodity. If this sum is greater than one, the production will not be economically
justifiable symbolically:
n

a ij
 1  j  1, 2 , .... , n  and each a ij
is non-negative or zero.
i 1

The cost of the primary input (which is termed also as value added) needed in producing a unit of
th
the j commodity would be :

 n

1 

a
i 1
ij


where a ij
is are in value terms

Solutions of open model

Consider an economy consisting of “n” industries if producing sector 1 is to produce and output
just sufficient to meet the input requirements of the “n” industries as the final demand of the

exogenous sector, its output level  1

must satisfy the following equation.

6|Page
 1
 a 
11 1
 a 12 2
 ......  a  F1n n 1

 2
 a 
21 1
 a  22 2
 .......  a   F
2n n 2

 
 n
 a 
n1 1
 a  n2 2
 ......  a 
nn n
 F n

In matrix notation the above system can be written as:

  a11  
 1 
a 12
..... a n1

  1
F1 
 
  a21  
 2  a 22
..... a n2 
    F 2 
2
 
        
 
  a  
  n   n1 a n2 ann   n   F n 

    F
Where A is given input coefficient matrix while x and F are the vectors of output and final
demand of each producing sector.
     F
  ( I  )  F
1
   (   ) F

1
If     0 then,     exists, we can then estimate for either of the two matrices X

and F by assuming that one of them to be given exogenously.

Example 1 suppose and economy is made only by three sectors and we need to estimate the
output of each sector with the given input coefficient matrix and final demand as follows
(coefficient matrix given in value terms)
P Q R
0.3 0.4 0.2 100
A= 0.2 0 0.5 F  40 

0.1 0.3 0.1  50 

Here, it should be noted that 3 – column sums of A are (0.3 + 0.2 + 0.1 ) = 0.6, (0.4 + 0 +
0.3) = 0.7 and (0.2 + 0.5 + 0.1) = 0.8 which are less than 1 in each case. In other words, (1-

7|Page
0.6) = 0.4 (1-0.7) = 0.3 and (1-0.8) = 0.2 is the maximum amount of primary input needed for
producing a dollar worth of the three commodities (  , Q and R) Respectively.

0.3 0.4 0.2 100

Since   0.2 0 0.5 and F  40  in milliions of dollars
 
 0.1 0.3 0.1 50 

 0.7  0.4 0.2 

     0.2 0.1  0.5
  0.1  0.3 0.9 
1
Substituting these values in      F 1
we have

 
 1  0.7  0.4  0.2 1
100
    2   0.2 1  0.5 40 
 
 
  0.1  0.3 0.9  50 

 
  3

 0.7  0.4  0.2 1

0.75 0.42 0.40
1
But  0.2 1  0.5  0.23 0.61 0.39
 0.401  
  0.1  0.3 0.9  0.16 0.25 0.62

0.75 0.42 0.40 100

1 0.23 0.61 0.39 40 
Thus,  
0.401    
0.16 0.25 0.62 50 

1
Therefore  1

0.401
0.75(100)  0.24(40)  0.40(50)  270 million dollar
1
 2

0.401
0.23(100)  0.61(40)  0.39(50)   167 million dollar
1
 3

0.401
0.16(100)  0.25(40)  0.62(50)   142 million dollar
Note that we can also analyze the amount of primary input that would be needed to get the above
output – max in the three producing sectors.

8|Page
We have already found above that the maximum amount of primary inputs (in money terms)
required for producing “ a dollar worth” of commodities in three sectors are 0.4 , 0.3 and 0.2
Thus, the total primary input requirements of the three sectors would be

L 1 1
 0.4 x 270  108 million in the first sec tor

L
nd
2
 0.3 x 160  50.1 million in the 2 sec tor
2

L
rd
3
 0.2 x 140  28.4 millioin in the 3 sec tor
3

Therefore a total of 108 + 50.1 + 28.4 = 186.5 million primary input is used up for the
production of the above output mix.
Example 2 An economy produces Oil and Coal. The two commodities serve as intermediate
inputs in each other’s production. 0.4 tons of Oil and 0.7 tons of Coal are needed to produce a
tons of oil. Similarly, 0.1 tons of oil and 0.6 tons of coal are required to produce a ton of coal.
Beside to these intermediate inputs, 2 and 5 labour days are required to produce a ton of oil and
coal respectively. If the economy needs 100 tons of oil and 50 tons of coal, calculate the gross
output of the two commodities and the total labour required.
Solution: We can form the technology matrix as:
Oil Coal Final demand
Oil 0.4 0.1 50
Coal 0.7 0.6 100
Labour 5 2 -

0.4 0.1
Thus,   
0.7 0.6

0.6  0.1
   
 0.7 0.4

1 1 0.4 0.1
    
0.17 0.7 0.6
 
1 0.4 0.1 50  176.5 
    1   
  0.17 0.7 0.6 100
  558.8

 2
The total labour days required is given by:

9|Page
5 oil output  2 Coal output
 5  2
1 2

 5 x 176.5  2 x 558.8  2,000 labour days

4.5. Coefficient matrix and closed models.

We shall now examine if we will be able to estimate for X if the model is changed into closed
one. If the exogenous sector (final demand bill) of the open input – output model is absorbed in
to the system of endogenous sectors, the model would turn in to a closed one. In such a model
the final demand bill and primary inputs will not appear any more, other in their place, we shall
have the input requirements and output of this newly conceived industry, the “household
industry” producing the primary input labour. Final demand sector would now be considered as
none of endogenous sectors.

This newly conceived industry (of final demand bill) will also be assumed to have a fixed input
ratio as any other industry. In other words, the supply of primary input must now bear a fixed
proportion to final demand (i.e. consumption of this newly conceived industry). This will mean,
for example, that households will consume each commodity in fixed proportion to the labour
services they supply.

Looking at the problem in this particular way, it appears that the conversion of open model in to
a closed one should create any significant change in our analysis and solution because
disappearance of final demand means only an addition of one more homogeneous equation to
already existing set of n homogeneous equations.

Let us assume that there are four industries only – including the new one (of final demand)
designated by subscript 0 We shall therefore , have the following set of equations:

 0
 a 
00 0
 a  a 
01 1 02 2
 a 
03 3

 1
a 10 0
 a   a 
11 1 12 2
 a 
13 3

 2
 a 20 0
 a   a 
21 1 22 2
 a 
23 3

 3
 a 30 0
 a  a 
31 1 32 2
 a 
33 3

This will be reduced into a homogeneous equation system as:

10 | P a g e
  
1  a00  a01  a02  a03   0 0 
    0 
 a10 1  a11  a12  a13 
 1
    
  0 
 a20 a 21
1  a22  a23 
 2  
 (1  a33) 0 
 a30 a  a32
 3
31
 

   X  0
Since, the 4 rows of the input coefficient matrix happen to be linearly dependent;    will

turn out to be zero. Hence the solution is indeterminate.

This means that in a closed model no unique output – mix of each sector exists. We can at most
determine the output levels of endogenous sectors in proportion to one another. But cannot fix
their absolute levels unless additional information is made available exogenously.
4.6. The howkins - simon conditions /The viability of the system/
Many a time input – output matrix solution may give outputs expressed by negative numbers. If
our solution gives negative outputs it means that more than one tone (or any unit ) of that product
is used up in the production of every one tone (unit ) of that product which is definitely an
unrealistic situation such a system is not viable system.
The howkin – simon conditions guard against such eventualities.

Given our basic equation     

1
F , In order that this does not give negative numbers

as a solution, the matrix    which infect is the Leontief matrix should be such that
a) the determinant of the matrix must always be positive
b) The diagonal elements: 1  a 11
) , (1 a 22
) , (1  a 
33
,    (1  a nn
be

positive or, in other words, elements a 11

, a ,a
22 33
 a nn
should all be less than one.

Thus, one unit of output of any sector should use not more than one unit of its own output. Such
conditions are called the Hawkins –simon conditions.
0.8 0.2 
Example: suppose A =  Test is whether the system is viable not.
0.9 0.7

0.2  0.2
Solution     
 0.9 0.3

11 | P a g e
Here the value of the determinant of    will be -0.12

Which is less than zero, though the diagonal elements in    are all positive. As such the
H - s conditions are not satisfied
Thus, No solution will be possible in this case
Example: The following inter– industry transaction table was constructed for and economy with
two industrial sectors.
Final
Industry I II consumption Total
I 500 1,600 400 2,500
II 1750 1,600 4,650 8,000
Labour 250 4,800 __ 5,050
Total 2,500 8,000 5050 15,550

Construct technology coefficient matrix showing direct requirements. Does a solution exist for
the system? If yes, calculate the total level of output from each industry.
Solution Technology matrix showing direct requirements per unit of output is obtained by
dividing each input by the total output of the sector. That is

 500
a 
11
 11

2,500
 0.2
1

 1,600
a  
12
12

8,000
 0. 2
2

 1750
a  
21
21

2500
 0. 7
1

 1600
a  
22
22

8000
 0.2
2

Hence the technology matrix is given by:

Industry I II

I 0.2 0.2
II 0.7 0.2

12 | P a g e
 Labour 0.1 0.6

1 0 0.2 0.2
A =  _
0 1  0.7 0.2

0.8  0.2
Thus,      0.64  0.14
 0.7 0.8
= 0.50
Since    is positive and all elements of the principal diagonal of    are all positive,

the Hawkins simon conditions are satisfied. Hence, the empirical system has a solution.

To find the total outputs of each industry let  1

and  2
be the total outputs from industry I

and II respectively.
  F1 
Then,      
1 1
 
   F 2 
 2
0.8  0.2 1 400 
= 
 0.7 0.2 4650
 
1 0.2 0.2 400 
0.5 0.7 0.8 4650

0.2 0.2  4000

= 2
0.7 0.8  4650
 
1010
= 2 
4000
2020
=  
8000
Thus, the total output from industry I is 2020 units and that of industry II is 8,000 units.

13 | P a g e