3.4.

The Input- Output (I-O) Model

1
A much more sophisticated approach to
development planning is to use some variant of
the interindustry or input-output model,
input-output analysis is a technique for
studying inter-industry connections and
interdependencies in the entire economy, where
one industry’s input equals another industry’s
output.
An input-output model represents national or
regional economic accounting, which records how
industries trade with each other as well as how
they produce for consumption and investment .
2
In I-O model the activities of the major industrial
sectors of the economy are interrelated by a set of
simultaneous algebraic equations expressing the
specific production processes
All industries are viewed both as producers of
outputs and users of inputs from other industries.
For example, the agricultural sector is both a
producer of output (e.g., wheat) and a user of
inputs from, say, the manufacturing sector
(e.g., machinery, fertilizer).

3
Thus, direct and indirect repercussions of
planned changes in the demand for the products
of any one industry on output, employment, and
imports of all other industries can be traced
throughout the entire economy in complex web
of economic interdependence.
 Given the planned output targets for each
sector of the economy, the interindustry model
(I-O Model) can be used to determine
intermediate material, import, labor, and capital
requirements with the result that a
comprehensive economic plan with mutually
consistent production levels and resource
requirements can, in theory, be constructed
4
 3.4.1. History and Development of the Model
The model was developed by Wassily Leontief, a Russian-born
economist, and he was awarded the 1973 Nobel Prize in
Economic Sciences
The I-O tracks intermediate purchases (industry to industry) and
final purchases (industry to households and industry to
government).
It allows researchers to measure the following
 Changes in final demand on the economy
 Total multiplier effects including employment, earnings, and output
 Inter-industry linkages.

5
While some investigators use I-O to analyze inter-industry
linkages, the majority of regional analysts– including those in
state agencies and local governments– use the model to report on
the multipliers.
3.4.2 .Assumptions and Mathematical Formulae
 The basic assumptions of the IO analyses:
 No substitution takes place between the inputs to produce a
given unit of output and the input coefficients are constant.
 The linear input functions imply that the marginal input
coefficients are equal to the average.
 Joint products are ruled out, i.e. each industry produces only
one commodity and each commodity is produced by only one
industry.
 External economies are ruled out. There are no external
economies or production diseconomies.
6
 A company’s total production is used as input by
another company or the final demand section.
Production is followed according to the law of
constant returns to scale
In the economic area, the level of technological
progress remains constant, so the input coefficients
remain constant.
 In the IO analysis we need to have the following:
First, an inter-industry transactions table is needed, which is termed
as X matrix
Second, that table is converted into a direct requirements table ,
which is called [A] matrix and,
 Finally a total requirements table, termed as [I-A]-1 matrix or
Multiplier or M-matrix.
7
The first row shows, the agriculture sector produces 500 units
of outputs in total, of which 320 units flow into the
manufacturing sector as inputs for its production. 100 units are
delivered to households directly as final demand, and the
remaining 80 units are consumed by the agriculture sector itself
as fodder and seeds
The 2nd row shows how outpus the manufacturing sector flow.
60 units flow to the agriculture sector as. 40 units flow back to
the manufacturing sector to support further production
8 processes. The remaining 160 units are consumed directly by
 Production Functions and the Input–Output
Model
a) Matrix 1: Transactions Matrix (X matrix)
It contains data on the flow of goods and services in the
economy.
It shows how the total output of the ith industry or sector of
the economy ( i.e , Xi) is distributed to all other sectors as
input (i.e Xij, where i=j and i= 1, 2…n, and j= 1, 2, …n,
with n number of non-final demand sectors in the economy)
and to the final demand sector (Di).

9
Production Functions and the Input–Output Model
a) Matrix 1: Transactions Matrix (X matrix
The transaction matrix can be represented by the following

Where: Xi = Gross output of sector i

Xij = Output of sector i used as input or intermediate input
by sector j.
Di = the final demand for the product of sector I
•It contains data on the flow of goods and services in the
10 economy.
•
 • It shows how the total output of the ith industry or
sector of the economy ( i.e , Xi) is distributed to all
other sectors as input (i.e Xij, where i=j and i= 1,
2…n, and j= 1, 2, …n, with n number of non-final
demand sectors in the economy) and to the final

.”

11
 b) The Technical coefficient matrix (Direct
requirements matrix) or A Matrix
The transactions matrix (X-Matrix) is transformed into a table of
direct requirement coefficients or called the technical coefficients
matrix ( A-Matrix).
The A-matrix shows the distribution of inputs (xij) per unit of
output E(xij). Each coefficient (aij) represents the amount one
industry purchases directly from another industry per dollars worth
of output.
Let there be n industries in the economy.
The input-output table in the form of the matrix A = [aij] would
state the input coefficients.

12
 Each column shows the necessary input for
producing one unit of the output of a certain industry.
If any element in the matrix is zero, it shows that the
input demand is zero.
The input or technical coefficients can be written as:

Where: Xj = the total output of the jth sector

Xij= the output of the ith industry used as input in the jth
industry

13
 Given Xij = input of aluminum (i) bought by aircraft
producers (j) last year
Xj – total aircraft production last year
Xij/xj = aij :
This ratio is called a technical coefficient; the terms input–output
coefficient and direct input coefficient are also often used.
For example, if X12 =Birr 300 and X2 =Birr15, 000 =
sector 4 used Birr 300 of goods from sector 1 in producing
$15,000 of sector 2 output), a12 =X12/X2 = Birr 300/Birr
15, 000=0.02. Since a14 is actually $0.02/$1, the 0.02 is
interpreted as the “dollars’ worth of inputs from sector 1
per dollar’s worth of output of sector 2 .

14
 For example, if
X12 =Birr 300 and
X2 =Birr15, 000 =

Sector 2used Birr 300 of goods from sector 1 in producing

$15,000 of sector 2 output),
a12 =X12/X2 = Birr 300/Birr 15, 000=0.02.

Since a12 is actually Birr 0.02/Birr1, the 0.02 is

interpreted as the “Birr’ worth of inputs from sector 1 per
Birr,s worth of output of sector 2.

15

:
The technical coefficients or direct requirements table is
usually given in the following matrix
Output
Input 1 2  n
1  a11 a12 ... a1n 
a a22 ... a2 n 
2  21 
A  . . 
 
. . 
. . 
 
n 
 an1 an 2  ann  
However, the Input-Output table in the above form considers
only inter-industry flows and ignores final demand.
An 'open' Input-Output table can be easily constructed where a
final demand for the product of each industry is included.
Note that corresponding to the demand for the products, the I-0
table should now be expanded to include supplies of primary
inputs
16
For example, if the household sector's final demand for
output is now included it is also necessary to include the
labor supplied by the households as inputs.
It is now obvious that, because of the supply of labor inputs,
the sum of the elements in each column of the matrix A will
be less than one because in the absence of primary input
costs (e.g. labor supply) the sum of each element in any
column will be exactly equal to one. Thus:
n


i 1
aij  1 j 1, 2, ..., n

It follows that the value of primary input required to

produce one unit of the jth good is given by
n
1 
i 1
aij
17
Now for industry 1, to produce enough output to cater for the
final demand plus the input demand of n industries, the following
equation must hold:
x1  a11 x1  a12 x 2  . . . ..  a1n x n  D1
or
1  a11 x1  a12 x 2  . . .  a1n x n  D1
where D1 is the final demand for the output of industry 1.
 Similarly, for the second industry, the equation can be set out as
below:
 a21 x1  1  a22 x2  .., .  a2 n xn  D2
Thus
 an1 x1  an 2 x2  ...  1  ann xn  Dn
18
 Matrix equations
We can represent Input -output matrices symbolically as follows:
X = AX+D
X – AX = D
IX – AX = D
(I – A)X = D
if the inverse of (I – A) exists.
The system of equations can be written in the following matrix
form:
I  A X  D

1  a11  a12  a1n   x1   D1 

     
  a21 1  a22  a2 n   x2   D2 
. . .  .  . 
      
. . .  .  . 
     
 .  .   . 
 a  an 2   ann  x  D 
 n1   2  2

Where [I-A] is called as the Leontief Matrix and also known as the
intermediate matrix.

20
 The identity matrix I is given by
1 0 0 0
 
0 1 0 0
0 0 1 0
 
I  . 
. 
 
. 
 
0 0 0 1 

That is, all the elements in the principal diagonal are 1 and all other
elements are zero.
Now solving for X we have:
X  I  A
1
D
 where, [I-A]-1 is the inverse of the Leontief Matrix and called as the
total requirements matrix
 The rule for matrix inversion, i.e. [I-A] -1 can be given as:
1
[ I  A] 1  [ I  A] *
IA
21 where, [ I  A] * is the Adjo int matrix of [ I  A]
The following example can be given. Let
1  a11 a12 
[ I  A ]   

a
 21 1  a 22 

Then
I  A  (1  a11 )(1  a22 )  a12 a21

and
1  a 22  a12 
[ I  A] *   


 21a 1  a11 

1 1  a 22  a12 
[ I  A] 1
  
IA  a  (1  a11 ) 
 21

 1  a22  a12 
 
 I  A I  A 
 
 a21 1  a11 
 
22  I  A I  A 
 
A numerical example. To simplify the analysis we are concerned with
only two sectors, agriculture (X1) and textiles (X2). Let the IO table be
given as:
Agriculture Textiles
Agriculture 0.6 0.2
Textiles 0.4 0.3
Then A=  0.6 0.2 
 
 04 0.3 
Now 1 0   0.6 0.2   0.4  0.2 
I  A         
 0 1  0.4 0.3    0.4 0.7 
I  A  (0.4 X 0.7)  ( 0.2 X  0.4)
 0.28  0.08
 0 .2

 0.7 0.2 
I  A  
*

 0.4 0.4 
 0.7 / 0.2 0.2 / 0.2   3.5 1
I  A  
1
   
24  0.4 / 0.2 0.4 / 0.2   2 2
 Let the final demand be given by
10 
D
5 
 

 Recall that X  I  A
1
D

 3 12 1  10   40 
 Thus we have X        
2 2 5   30 
Thus the agricultural sector (X 1) would produce 40 units
and the textile sector (X2) would produce 30 units.

25
3.5.3 Numerical Example
Let’s assume a simple country economy. All things being equal,
this country has seven industries. These industries sell goods and
services to each other and purchase goods and services from each
other.
Let’s further assume that the country’s Bureau of Economic
Analysis has provided us with a transactions matrix. One of the
fundamental rules is that total inputs (xi) should equal total
outputs (xj). In other words, just as in accounting, the books have
to be balanced.
The following table represents a seven by seven transaction
matrix:

26
  X ij  Di  Xi
where , i  1,2,3    7 and j  1,2,3    7
1 2 3 4 5 6 7

Ag Mi Co Ma Tr T S Final Total
Demand Output
Agriculture 51949 80 1415 91220 4619 11187 348 35684 196502
(Ag)
Mining (Mi) 369 18346 4969 170902 1008 51499 3646 27477 223262

Construction 1534 470 826 20018 12473 66388 11115 381933 494757
(Co)
Manufacturing 3672 15700 176203 865110 79418 211830 8678 1016471 2377082
(Ma)
Transportation 10360 4342 56747 150767 66123 71618 5260 693175 1058392
(Tr)
Trade (T) 14936 20698 47070 180771 172066 468030 15100 1347464 2266135

Services (S) 451 1035 1645 25133 16832 30942 2062 398453 476553

Value Added 80150 162611 205881 906224 705952 1355091 429895 0 3845804
(VAT)
Imports 33081 -2 0 0 -99 -449 449 101 33081

Xi 196502 223280 494756 2410145 1058392 2266136 476553 3845804 10971568
27
The table above represents transactions within a stable
economy,
but what if we introduced a change, e.g., new capital
investment. How would that affect the country economy?
To answer the question, we need to inspect the basic structure
of the economy, its technical coefficients or per unit costs.
The following table represents the technical coefficient matrix
A  a  ij

X ij
where , aij  ; i  1, 2,3 7; j  1, 2,3 7
X j

1 2 3 4 5 6 7
Agricultur e (1)  0.2644 0.0004 0.0029 0.0378 0.0044 0.0049 0.0007 
Mining (2)  0.0019 0.0822 0.0100 0.0709 0.0010 0.0277 0.0077 
 
Constructi on (3)  0.0078 0.0021 0.0017 0.0083 0.0118 0.0293 0.0233 
 
A  Manufactur ing (4)  0.0187 0.0703 0.3561 0.3589 0.0750 0.0935 0.0182 
Transporta tion (5)  0.0527 0.0194 0.1147 0.0626 0.0625 0.0316 0.0110 
 
Trade (6)  0.0760 0.0927 0.0951 0.0750 0.1626 0.2065 0.0317 
Services (7)  0.0023 0.0046 0.0033 0.0104 0.0159 0.0137 0.0043 
28  
Next, we must create the total requirements matrix. To do that,
however, we need to run through a few intermediate steps.
First, create a seven by seven identity matrix. Look along the
main diagonal of the matrix below: all those cells have a
number one.
1 0 0 0 0 0 0
0 1 0 0 0 0 0
 
0 0 1 0 0 0 0
 
I  0 0 0 1 0 0 0
0 0 0 0 1 0 0
 
0 0 0 0 0 1 0
0 0 0 0 0 0 1
 

Second, subtract the technical coefficients matrix (A) from the

identity matrix (I). The result is the matrix below.

29
  0.7356  0.0004  0.0029  0.0378  0.0044  0.0049  0.0007 
  0.0019 0.9178  0.0100  0.0709  0.0010  0.0277  0.0077 

  0.0078  0.0021 0.9983  0.0083  0.0118  0.0293  0.0233 
 
I  A     0.0187  0.0703  0.3561 0.6411  0.0750  0.0935  0.0182 
  0.0527  0.0194  0.1147  0.0626 0.9375  0.0316  0.0110 
 
  0.0760  0.0927  0.0951  0.0750  0.1626 0.7935  0.0317 
  0.0023  0.0046  0.0033  0.0104  0.0159  0.0137 0.9957 

Now, we create the Leontief Inverse or total indirect
requirements matrix, also known as the matrix of multipliers.
 1.3655 0.0098 0.0389 0.0865 0.0175 0.0212 0.0044 
 0.0137 1.1055 0.0657 0.1323 0.0217 0.0509 0.0143 

 0.0174 0.0089 1.0166 0.0227 0.0224 0.0419 0.0259 
 
 I  A  1   0.0853 0.1540 0.6278 1.6354 0.1800 0.2280 0.0551 
 0.0904 0.0402 0.1777 0.1268 1.0915 0.0670 0.0211 
 
 0.1614 0.1543 0.2297 0.2079 0.2484 1.3095 0.0549 
 0.0078 0.0096 0.0164 0.0229 0.0229 0.0219 1.0062 


What exactly does the Leontief Inverse tell us?

It tells how $1 dollar spent by an industry impacts other
30
industries in the economy.
For example, in our fictitious country, $1 spent by Industry 2
increases activity in Industry 2 by almost $1.11; $1 spent by
Industry 2 increases activity in Industry 4 by $0.15 cents.
These unique linkage relationships while interesting do not
answer the macro question: What impact does a final demand
change in Industry i have on the whole economy?
Suppose that our example country embarks on a $2.75 million
manufacturing revitalization program, how will that change in
final demands affect the economy? To determine the answer, we
multiply the [I-A]-1 matrix by the ‘change (ΔD)’ matrix.
I  A  1
X D  Im pact
 1.3655 0.0098 0.0389 0.0865 0.0175 0.0212 0.0044   0   $23775 
 0.0137 1.1055 0.0657 0.1323 0.0217 0.0509 0.0143   0   $363825 
   
 0.0174 0.0089 1.0166 0.0227 0.0224 0.0419 0.0259   0   $62425 
     
 0.0853 0.1540 0.6278 1.6354 0.1800 0.2280 0.0551  X $2750000  $4497350
 0.0904 0.0402 0.1777 0.1268 1.0915 0.0670 0.0211   0   $348700 
     
 031
.1614 0.1543 0.2297 0.2079 0.2484 1.3095 0.0549   0   $571725 
 0.0078 0.0096 0.0164 0.0229 0.0229 0.0219 1.0062   0   $62975 
   
The table above shows us the potential increase each industry in
the country will experience because $2.75 million dollars was
introduced to revitalize the manufacturing industry—industry #4.
The country overall will benefit by $6.14 million— sum of the
impact column.
The manufacturing multiplier, for this country, is 2.23– divide
the total impact by the direct effect.
Alternatively, to calculate an industry multiplier, add each of the
indirect requirements coefficients, found in the Total
Requirements Table (matrix), down a column.

32
 Types of Impacts in Input-Output
Analysis
A. Direct effects
 The revenues, jobs, and wages that a new
business or expanding business brings into the
local economy-- or removes from the economy
as the case might be.
 The impacts of a change in final demand on the
consumption of the directly associated inputs.
 For example, building a dam requires steel,
concrete, workforce, and construction machinery.
It thus has a direct impact on these inputs.
33
 B. Indirect effects
 Any business expansion or new entry into a market will lead
that business to make purchases from and/or sales to local
firms. Because of new demand, local firms are likely to
create some number of new jobs, increase wages and
revenues. All this, in turn, will have an additional impact on
the overall economy.
C. Induced effects
While direct effects and indirect effects measure the impacts
of business-to-business interactions, induced effects are
specific to the behavior of the labor force. Thus, employees
of the new business and the related businesses will spend
their earnings in the local economy to purchase items such
as
34
food, transportation, housing, medical, etc.. This spending
will have additional impacts on the overall economy.
3.5.4 Limitations of the Model
National benchmark input-output models are constructed once
every five Where: Xj = the total output of the jth sector
years and the data may not be available to the public for another
5 years.
Changes in the economic structure that occurred between the end
of data collection and end of calendar year are not recorded.
Thus, we have problems of currency and accuracy.
Input-output models are also static. That is to say, an IO model is
a snapshot of the economy at one point in time.
Specific limitations to the input-output model’s accuracy include:
Constant coefficients
Linearity
Sector homogeneity

35 No capacity constraints
Constant coefficients imply that advances in production
technology, new inventions, import substitution, changes in
consumer patterns of demand, and the increase or decrease of
relative prices do not alter aij.

Despite these limitations, the input-output model is perhaps one

of the more powerful descriptive tools available to the regional
analyst.

36