# Detailed Key Assumptions of IMPLAN & Input-Output Analysis

## INTRODUCTION

Input-Output (I-O) analysis is a means of examining inter-industry relationships within an economy. It captures all monetary market transactions between industries in a given time period. The resulting mathematical formulae allow for examinations of the effects of a change in one or several economic activities on an entire economy (known as economic impact analysis).

IMPLAN expands upon the traditional I-O approach to also include transactions between industries and institutions¹ and between institutions themselves, thereby capturing all monetary market transactions in a given time period. IMPLAN can thus more accurately be described as a Social Account Matrix (SAM) model, though the terms I-O and SAM are often used interchangeably.

Although IMPLAN provides a framework to conduct an analysis of economic impacts, each stage of an analysis should be carefully scrutinized to make sure it is logical. Procedures and assumptions need to be validated. Please review IMPLAN and Input-Output analysis' assumptions.

## ASSUMPTIONS

### Constant Returns to Scale

The same quantity of inputs is needed per unit of Output, regardless of the level of production. In other words, if Output increases by 10%, input requirements will also increase by 10%.

### Fixed Input Structure

This structure assumes that changes in the economy will affect the Industry's Output level but not the mix of Commodities and services it requires to produce that Output. This means that the same recipe of inputs will always be used to create the Output unless changes to the production function are made by the user.

### Industry Homogeneity

I-O models assume that all firms within an industry are characterized by a common production process. In IMPLAN, edits can be made to the production function of an industry in order to model a distinct firm.

### No Supply Constraints

I-O assumes there are no restrictions to inputs, raw materials, and employment and assumes there is enough to produce an unlimited amount of product. It is up to the user to decide whether this is a reasonable assumption for their study area and analysis, especially when dealing with large-scale impacts.

### Fixed Technology

An Industry, and the production of Commodities, uses the same technology to produce each of its products. In other words, an Industry's Leontief Production Function is a weighted average of the inputs required to produce the primary product and each of the byproducts, weighted by the Output of each of the products.

### Constant Byproduct Coefficients

As a requirement of the technology assumption, Industry byproduct coefficients are constant. An Industry will always produce the same mix of Commodities regardless of the level of production.

### The Model is Static

No price changes are built in IMPLAN and the underlying data and relationships are not affected by impact runs. Input-Output models do not account for general equilibrium effects such as offsetting gains or losses in other Industries or geographies or the diversion of funds from other projects.

In Input-Output models, Type I multipliers measure only the backward linkages, also known as upstream effects. Input-Output analysis does not look at forward linkages in terms of how an Industry’s production is used as an input for other production or for final use, also known as downstream effects.

### Time Dimension

The length of time that it takes for the economy to settle at its new equilibrium after an initial change in economic activity is unclear because time is not explicitly included in regional I-O models.

## PRINCIPLES OF INPUT-OUTPUT

There are three fundamental principles underlying the I-O accounts (Horowitz & Planting, 2009).

### CONSISTENCY PRINCIPLE

Under this principle, the data compiled from one source are comparable with the data compiled from another source. For example, in accordance with this principle, the estimates shown in the I-O accounts should be consistent with the underlying source data and with the estimates shown in the national accounts. In the United States, NAICS provides a consistent basis for classification that enables comparisons across the broad range of economic statistics.

### HOMOGENEITY PRINCIPLE

Under this principle, each Industry's Output is produced with a unique set of inputs or a unique production function.

### PROPORTIONALITY PRINCIPLE

Under this principle, all inputs consumed by an Industry are a linear function of the level of Output, that is the inputs consumed vary in direct proportion to Output and there are no economies of scale.

## Other Considerations

### IMPLAN Disclaimer

IMPLAN is a regional economic analysis software application that is designed to estimate the impact or ripple effect (specifically backward linkages) of a given economic activity within a specific geographic area through the implementation of its Input-Output and Social Accounting Matrix. Studies, results, and reports that rely on IMPLAN data or application are limited by the researcher’s assumptions concerning the subject or event being modeled. Studies are in no way endorsed or verified by IMPLAN Group, LLC unless otherwise stated by a representative of IMPLAN.

IMPLAN provides the estimated Indirect and Induced Effects of the given economic activity as defined by the user's inputs. Some Direct Effects may be estimated by IMPLAN when such information is not specified by the user. While IMPLAN is an excellent tool for its designed purposes, it is the responsibility of analysts using IMPLAN to be sure inputs are defined appropriately and to be aware of the assumptions within any I-O and Social Accounting Matrix Model.

Analysts should evaluate the logistical feasibility of a business outside of IMPLAN, as IMPLAN models cannot determine whether it will be financially successful. Additionally, I-O models do not account for forward linkages, nor do I-O models account for offsetting effects such as cannibalization of other existing businesses, diverting funds used for the project from other potential or existing projects, etc.  It falls upon the analyst to take such possible countervailing or offsetting effects into account or to note the omission of such possible effects from the analysis.

### VALUES THAT MUST BE CONVERTED BEFORE IMPLEMENTING IMPLAN ANALYSIS:

• Price changes
• Policy Changes
• Socio-political Impacts
• Environmental Impacts
• Feasibility Analysis
• Net Effects of Innovation & Transitions
• Opportunity Costs

These are a few examples of complex issues that require special framing for IMPLAN analysis. All these cases have associated economic elements (forest management, construction, healthcare costs, social works programs, etc.), and their aspects related to spending and production can be modeled using IMPLAN. Likewise, pairing IMPLAN with carbon emission reports allows analysts to create estimates of output-to-carbon emission values. However, the environmental impacts of those emissions are based on the analyst's assumptions. Therefore, two different views could produce radically different results.

### WHY ISN'T THERE ANY FORECASTING?

IMPLAN estimates impacts assuming that the relationships of the current data year are maintained. Thus, IMPLAN datasets are a snapshot in time of the economy. While estimates of economic activity related to specific demand changes and their associated supply linkages can be estimated with IMPLAN, the software cannot predict what the total employment in a state will be five years from now. The economy of even a small region is constantly in flux, affected by decisions made in businesses and households, by policies, and even environmental factors that can contribute to whether a region thrives, stagnates, or dwindles. To make a projection of what the economy will look like five years from now, it would be necessary to predict all the demands for consumption and know all the new commodities and technologies that will be available. Local availability of resources to meet that demand would also need to be known.

## CITATIONS

Adams, A.A. & Stewart, I.G. (1956). Input-Output Analysis: An Application. The Economic Journal, 66 (263), 442-454.

Bess R. & Ambargis, Z.O. (2011). Input-Output Models for Impact Analysis: Suggestions for Practitioners Using RIMS II Multipliers. Presented at the 50th Southern Regional Science Association Conference,  New Orleans, Louisiana. https://www.bea.gov/system/files/papers/WP2012-3.pdf

Christ, C.F. (1955). A Review of Input-Output Analysis. In Input-Output Analysis: An Appraisal (pp. 137-182). Princeton University Press.

Guo, J., Lawson, A.M., & Planting, M.A. (2002). From Make-Use to Symmetric I-O Tables: An Assessment of Alternative Technology Assumptions. Presented at the 14th International Conference on Input-Output Techniques, Montreal, Canada. https://www.bea.gov/system/files/papers/WP2002-3.pdf

Horowitz, K.J. & Planting, M.A. (2009). Concepts and Methods of the U.S. Input-Output Accounts. Bureau of Economic Analysis, US Department of Commerce. https://www.bea.gov/sites/default/files/methodologies/IOmanual_092906.pdf

Miller, R.E. and P.D. Blair. (2009). Input-Output Analysis: Foundations and Extensions, Second Edition. New York: Cambridge University Press.

1. In IMPLAN, institutions include Households (broken down into nine income categories), Administrative Government, Enterprises (basically corporate profits), Capital, Inventory, and Foreign Trade.

Written August 30, 2023