What Is Input-Output Analysis? (Econ Model)
Input-output analysis is an economic model that shows how industries buy from and sell to one another. It estimates how much total production may be needed when demand for a product changes. The Leontief model uses a transactions table and matrix algebra to count both direct production and the inputs required throughout the supply chain.
If you are new to the model, the symbols can make a clear idea seem harder than it is. Start with a simple question: if people or businesses want more of one product, what else must industries produce to supply it?
Input-output analysis helps answer that question by mapping purchases between industries. It can support economic impact studies, policy work, and industry planning. But its results depend on the data and assumptions used. They are conditional estimates, not proof of cause or exact forecasts.
Input-Output Analysis: Purpose, Scope, and the Leontief Model
Input-output analysis describes the flow of goods and services between industries in an economy. Its basic Leontief model links total industry production to two things: production used as inputs by other industries, and final demand from buyers outside those production chains.
Think of a bakery that needs flour, power, and delivery services. Those suppliers, in turn, need inputs of their own. A change in demand for bread can therefore lead to production needs beyond the bakery itself. Input-output analysis tracks these connected requirements.
In the open Leontief model, “open” means final demand is treated as coming from outside the model. It may include household purchases, government spending, investment, or exports, depending on how the data define it. The model does not explain how that demand arose.
The core equation is:
[ x=Ax+d ]
Here, (x) is the vector of each industry’s gross output, or total production. (A) is the direct-requirements matrix, which records the inputs needed to produce output. And (d) is the vector of final demand.
Rearrange the equation:
[ (I-A)x=d ]
If (I-A) is invertible, then:
[ x=(I-A)^{-1}d ]
The matrix ((I-A)^{-1}) is called the Leontief inverse. It includes the production needed for final demand and the inputs that suppliers need from one another. That is why its results include direct and indirect production requirements.
The model is called “open” because some buyers are outside its industry network, not because it covers every part of the economy. Its boundary matters. A national table, for example, may treat imports differently from a regional table.
Core Curriculum: Transactions Tables, Technical Coefficients, and Matrix Requirements
A transactions table records sales from one industry to another, often alongside sales to final users. From these flows, analysts calculate technical coefficients: the amount of each industry’s input needed for one unit of another industry’s output.
Read the transactions table before building the matrix
A transactions table, also called an inter-industry table, lays out who supplies what to whom. Rows usually show the supplying industry, while columns show the industry buying or using those goods and services. This row-and-column convention is common, but check the table’s notes because formats can differ.
Let (z_{ij}) mean the value of sales from supplying industry (i) to purchasing industry (j). Let (x_j) be the total output of purchasing industry (j). Then the technical coefficient is:
[ a_{ij}=\frac{z_{ij}}{x_j} ]
In plain English, (a_{ij}) tells you how much input from industry (i) is required to produce one unit of industry (j)’s output. For example, if (a_{ij}=0.10), the model records 10 cents of industry (i)’s input for each dollar of industry (j)’s output, when values use the same dollar basis.
| Item | Meaning | Check before using it |
|---|---|---|
| (z_{ij}) | Sales from supplier (i) to producer (j) | Confirm row and column labels |
| (x_j) | Total output of producer (j) | Use the same period and valuation basis |
| (a_{ij}) | Input from (i) per unit of output from (j) | Divide by the output of column (j) |
| (d_j) | Final demand for output from industry (j) | Check which buyer groups are included |
| (x_j) in the solution | Total output needed from industry (j) | Distinguish it from final demand |
The table’s values must use a consistent basis. Analysts should identify the industries, year, time period, and units, such as dollars or physical quantities. They should also check whether values are reported in current prices or adjusted for price changes.
Understand what the Leontief inverse adds
The direct-requirements matrix (A) shows the immediate inputs needed for production. The Leontief inverse adds the further production needed along the supply chain.
For example, producing a product may require electricity. Generating that electricity may require fuel and maintenance services. The inverse accounts for those linked requirements within the industries and assumptions included in the table. It does not mean every effect on people, prices, or the wider economy has been included.
The matrices need compatible dimensions. If a model has (n) industries, then (A), (I), and ((I-A)^{-1}) are (n \times n) matrices. The vectors (x) and (d) each have (n) entries. Matching industry order in every matrix and vector is essential.
Working Through the Model: Calculation, Interpretation, and Common Errors
To calculate an output requirement, define the industry categories and final-demand change, then solve ((I-A)x=d). A small example shows how intermediate purchases increase total production beyond the initial demand change.
A two-industry example
Suppose an illustrative economy has two industries: agriculture and manufacturing. Use the row-supplier, column-user convention, and assume all values share the same unit and period. Let the direct-requirements matrix be:
[ A= \begin{bmatrix} 0.20 & 0.10\ 0.30 & 0.20 \end{bmatrix} ]
The first column says that producing one unit of agriculture requires 0.20 units from agriculture and 0.30 from manufacturing. The second column says that producing one unit of manufacturing requires 0.10 units from agriculture and 0.20 from manufacturing.
Now assume final demand rises by 100 units for agriculture and does not change for manufacturing:
[ d= \begin{bmatrix} 100\ 0 \end{bmatrix} ]
The model is:
[ \begin{bmatrix} 0.80 & -0.10\ -0.30 & 0.80 \end{bmatrix} x= \begin{bmatrix} 100\ 0 \end{bmatrix} ]
Solving gives:
[ x\approx \begin{bmatrix} 131.15\ 49.18 \end{bmatrix} ]
So the model requires about 131.15 units of agricultural output and 49.18 units of manufacturing output. The direct final-demand increase is 100 units. The additional 80.33 units of gross production arise from the modeled production requirements across both industries.
These figures are illustrative, not observed economic data. The result depends on the chosen coefficients. It does not say that demand will necessarily rise by 100 units or that firms can produce the required amounts.
Check whether a solution exists
The inverse solution requires (I-A) to be nonsingular, meaning it has an inverse. A standard sufficient condition for a productive nonnegative model is that the spectral radius of (A), the largest absolute value of its eigenvalues, is less than 1.
For an introductory analysis, you may use software to check whether the model can be solved. Do not treat a successful calculation alone as proof that the data or assumptions make economic sense. Inspect the coefficients, units, and industry mapping too.
Avoid common interpretation errors
Several mistakes can make an output estimate misleading:
- Reversing the coefficient denominator: (a_{ij}) is divided by the output of industry (j), the purchasing industry in column (j).
- Confusing (A) with the Leontief inverse: (A) contains direct input needs; ((I-A)^{-1}) contains total direct and indirect requirements.
- Mixing units or years: A coefficient from one price basis or period may not fit a demand vector from another.
- Calling output “value added”: Gross output includes intermediate sales. It is not the same measure as value added or GDP.
- Treating the result as a forecast: The model calculates requirements under its assumptions. It does not establish what will happen.
Before reporting results, confirm matrix dimensions, coefficient meaning, data year, units, and source. Then state the assumptions and report sector-level results, not only a combined total.
Applications and Career Pathways: Economic Impact, Policy, and Industry Analysis
Input-output analysis is used to study how production needs connect across industries. Its applications include economic impact estimates, policy analysis, and supply-chain research. Each use depends on the model’s industry detail, data, and treatment of final demand.
Illustrative applications, not verified case studies
A regional analyst might estimate the production needed to meet an assumed increase in demand for a local service. A policy analyst might compare industry requirements under two final-demand scenarios. A company might use industry-level data to explore links between its sector and suppliers.
These are applications of the method, not claims about a specific project or measured result. A sound report states the demand change being tested, the data source and year, the model boundary, and what the output measure represents.
| Question | What the model can estimate | What it cannot establish by itself |
|---|---|---|
| What production is linked to a demand change? | Direct and indirect output requirements | Whether that demand change will occur |
| Which industries are connected? | Input needs recorded in the table | Every supplier relationship in real time |
| What might a policy scenario require? | Conditional output under stated assumptions | Full effects on prices, wages, or behavior |
| What does a regional estimate show? | Production linked to included local industries and data | All spending retained in the region, unless modeled |
Some studies also report induced effects, which refer to demand linked to income earned by workers and owners. The basic open Leontief model does not include these effects automatically. Analysts must add further modeling choices and clearly explain them. Adding induced effects without defining them can lead to double counting or overstatement.
Know the model’s limits
The basic model assumes fixed input proportions and constant technical coefficients. In other words, it assumes industries use the same input mix as output changes. It also does not, by itself, model price substitution, changing production methods, or supply constraints.
This matters when an industry lacks spare capacity, faces supply delays, or can switch to different inputs. In those situations, actual production and prices may respond differently from the model’s calculation. The result is best read as a conditional estimate: if the stated demand change occurs and the assumptions hold, the model indicates these production requirements.
Conclusion
Input-output analysis turns industry purchase-and-sale data into an estimate of the production tied to final demand. The direct-requirements matrix shows immediate inputs, while the Leontief inverse captures direct and indirect production needs. Read the result alongside its data, assumptions, and limits.
When you evaluate an analysis, ask what industries and year it covers, how final demand is defined, and whether the report separates gross output from other measures. Those checks help you understand what the numbers do and do not say.
Frequently Asked Questions
These short answers clarify the model’s main terms and uses. They also highlight the distinctions most likely to affect how you read an input-output result, from technical coefficients and multipliers to the model’s limits.
What is input-output analysis in economics?
It is a method for tracking purchases and sales between industries and estimating production requirements linked to final demand.
What is the Leontief model?
It is an input-output model written as (x=Ax+d), where total output meets both intermediate input needs and final demand.
What does the technical coefficient (a_{ij}) mean?
It measures the input from supplier industry (i) needed for one unit of output from industry (j), using a defined and consistent unit.
What is the Leontief inverse?
The matrix ((I-A)^{-1}) gives total direct and indirect output requirements for each unit of final demand, under the model’s assumptions.
Does input-output analysis predict the future?
No. It estimates requirements for a stated demand scenario. It does not prove that the scenario will happen or provide an unconditional forecast.
What is the difference between direct and indirect effects?
Direct requirements are immediate inputs for production. Indirect requirements are the further production needed to supply those inputs and related intermediate needs.
Does the basic model include induced effects?
No. Induced effects, such as demand linked to earned income, require additional modeling choices and are not part of the basic open Leontief model.
Why must (I-A) be invertible?
The standard inverse formula (x=(I-A)^{-1}d) only works when (I-A) has an inverse. Analysts should also check that the coefficients and data are suitable.
Is gross output the same as GDP?
No. Gross output includes intermediate transactions between industries. GDP measures value added, so the two totals are not interchangeable.
(This article was written by one of our staff writers, Nathan Sterling. Visit our Meet the Team page.)
