Input-Output Model and its Application
3.6 Structural limitations of input-output modeling
Indirect 0.18 0.22 0.35 Direct
1.00 0 0 Results of type I
output multiplier
Type I OP multp divided
Type I OP multp divided (2)
Impacts Agriculture Manufacturing Services
AG 1.18 0.22 0.35
Figure 3-10 Decomposing type I into direct and indirect impact.
Notes: AG, agricultural sector; multp, multiplier; OP, output.
(5) Type I income multiplier for agricultural sector(ab) /a (0.30.25)/0.31.87
Indirect 0.18 0.22 0.35 Direct
1.00 0 0 Results of type I
income multiplier
Type I OP multp divided
Type I OP multp divided (2)
Impacts Agriculture Manufacturing Services
AG 1.18 0.22 0.35
Income/output 0.3 0.23 0.45
Income/output 0.3 0.23 0.45
Direct income 0.30 0 0 0.30 (a)
Indirect income 0.05 0.05 0.16 0.26 (b) Direct income impacts
Indirect income impacts
(3)
(3) (4)
(4)
Figure 3-11 Type-I income multiplier calculation processes for the agricultural sector.
Notes: AG, agricultural sector; multp, multiplier; OP, output.
I N P U T-O U T P U T M O D E L A N D I T S A P P L I C AT I O N 67
and make you resilient to a series of possible criticism of its limitations, which will be described in the following sections.
3.6.1 No Assumption on supply constraints
I-O modeling assumes that the required input is always available without constraints in sup- ply. You can, for example, conduct a study of what would happen to a local economy if the
(5) Type I income multiplier for manufacturing sector(ab)/a (0.20.44)/0.22.97
Indirect 0.37 0.55 0.48 Direct
0 1.00 0 Results of type I
output multiplier
Type I OP multp divided
Type I OP multp divided (2)
Impacts Agriculture Manufacturing Services
MNF 0.371 1.55 0.48
Income/output 0.3 0.23 0.45
Income/output 0.3 0.23 0.45
Direct income 0 0.23 0 0.23 (a)
Indirect income 0.11 0.13 0.22 0.45 (b) Direct income impacts
Indirect income impacts
(3)
(3) (4)
(4)
(5) Type I income multiplier for srevices sector(ab) /a (0.450.23)/0.451.52
Indirect 0.12 0.21 0.33 Direct
0 0 1.00 Results of type I
output multiplier
Type I OP multp divided
Type I OP multp divided (2)
Impacts Agriculture Manufacturing Services
Serv 0.12 0.207 1.332
Income/output 0.3 0.23 0.45
Income/output 0.3 0.23 0.45
Direct income 0 0 0.45 0.45 (a)
Indirect income 0.04 0.05 0.15 0.23 (b) Direct income impacts
Indirect income impacts
(3)
(3) (4)
(4)
Figure 3-12 Type-I income multiplier calculation processes for the manufacturing and service sector.
Notes: emp, employment; MNF, manufacturing; Multp, multiplier; OP, output.
number of customers to a popular tourist destination – an ice-cream factory – were to triple in the following year. The model calculates that the required input of sugar, milk, and electric- ity will increase dramatically, which would stimulate the local economy through industrial interdependency. The model will not ask you whether there will be enough machines in the factory or enough parking spaces and seats for the tourists in that situation, because it will not make any assumption about capacity constraints.
Factory production in the study region can increase by 50% but I-O modeling tacitly assumes the existence of excess capacity, parking spaces, truck deck for delivery, and electricity, so that there would be no supply constraints to meet the new level of final demands. You can calculate the eco- nomic impact of increased tourist expenditure in Ghana, for example, assuming that there would be enough electricity supplied for hotel operations to meet the growing demand for hotel rooms. When the declining water level in the river threatens the amount of electricity generated by the hydraulic power generation plants, such a shortage of critical input creates a problem of supply constraints for hotel operations. These constraints, however, are not assumed in generic I-O modeling.
3.6.2 Constant return to scale
The required input can be put in relative terms as ratios, as we did when transforming the transaction table into the standardized A-matrix table. That relationship is assumed to con- tinue. If the input to the orange juice factory from the agricultural sector is 0.12, when the total output increases by $100 million, there should be an additional purchase of $12 million of agricultural product (oranges). If the increase in total output is of $100, there should be an additional purchase of oranges of $12. The relationship is assumed to remain linear.
For example, in a restaurant operation that is an independent outlet or part of a hotel, a certain amount of food is purchased as intermediate goods, and a certain amount of labor input is required to produce complete meals for the final demands from customers. The I-O is created based on the data collected at a certain point in time. One can also imagine the appearance of new businesses, such as half-cooked meals, or ready-made food ingredients. A manager now has the choice to purchase those ready-made ingredients, which will perhaps increase the proportion of the food cost, and possibly decrease the relative labor input, as the manager would not need as many prep-cooks. In the transactions table, there would be more purchase of manufactured goods in lieu of agricultural goods. Such changes are not included in the I-O framework, as you can understand by recalling the steps you took to create various matrices. It would not be an issue if only one or two firms in the industrial sector changed their purchasing behavior (input patterns). When there is a certain trend in the industrial sec- tor to change the purchasing patterns as a whole, the need may arise to update the I-O data.
3.6.3 Fixed commodity input structure
The fixed commodity input structure is similar to constant returns to scale, but the ratio of inputs is assumed to remain the same as it was during the observed study period, irrespective of changes in price of some items. In the real world, if the price of furniture from China
I N P U T-O U T P U T M O D E L A N D I T S A P P L I C AT I O N 69
becomes so attractive, a general manger of a hotel in the US may switch to purchases from China instead of those from a domestic manufacturing industry in West Virginia. If the price of beef increases for some reason, you may find fewer beef dishes at the restaurant. Increase in prices would not cause a huge technical problem for the I-O modelers if the change in prices of goods and services are equally distributed across all of society. That is often not the case, such as the surge of crude oil prices. In the I-O world, the exact mixture of intermediate goods will remain fixed despite short-term price fluctuations.
3.6.4 Homogeneous sector output
If the sector produces more than one commodity (this will be discussed later), the proportion of such multiple commodities productions will remain the same as in the study period. If an auto- mobile manufacturing factory in the region (the industry) produces more than one model of cars (the commodities), and I-O table recorded industry by commodities, the relative portion of com- modities (small cars, mid-size cars, large cars, huge trucks, gigantic sport utility vehicles) will not be assumed to change even when the gasoline prices increase later. As an example of the hospitality industry, let us consider a full-service hotel. They have a rooms department and food and beverage department as a core department to generate their products, clean rooms, and tasty meals . The proportion of those products is not assumed to change, which means that if the proportion was 70% and 30%, then when the total revenue at the hotel increased from $100 mil- lion to $150 million, the sales of the two departments are assumed to increase from $70 million and $30 million to $105 million to $45 million as the relative ratio of proportions are fixed.
In general, relatively small changes to the economy in question would pose less of a prob- lem, and changes to relatively larger economic region would create fewer problems in the impact analysis using the I-O framework. Being aware of those structural limitations, mainly derived from linear-modeling techniques, you may as well present your arguments in perspective with- out misleading audiences and readers. It is evident that these are not the all limitations of your particular study, as there may be errors that are more specific and applicable to your research.