An input-output table is a statistical table that shows, for a whole economy, how the output of each industry or product is used: as an input by other industries, or by final users such as households, government, investors and foreign buyers. Read across a row, it shows where an industry's sales go; read down a column, it shows what that industry buys to produce its output. The method was developed by the economist Wassily Leontief from the 1930s, work for which he received the Nobel memorial prize in economics in 1973.
Input-output tables are compiled by national statistical offices as part of the national accounts, following the System of National Accounts (SNA). In the United States the National Income and Product Accounts are supported by input-output (I-O) accounts published by the Bureau of Economic Analysis; in the United Kingdom the Office for National Statistics publishes analytical tables alongside the Blue Book. They are used to check the consistency of GDP estimates and to study how a change in one sector ripples through others.
Structure of the table
The core of a table is a square block of intermediate flows. Each cell records the value of output of the industry in the row that is bought as an input by the industry in the column. To the right sit columns of final demand, and below sit rows of value added: compensation of employees, operating surplus and taxes less subsidies on production. Imports may be shown as a separate row or split out by product.
Two accounting identities hold. For each row, total output equals intermediate sales plus final demand. For each column, total output equals intermediate purchases plus value added. Summed across the economy, total value added equals total final demand less imports, which is GDP measured two ways.
Supply and Use Tables
Most countries do not compile input-output tables directly. They first build supply and use tables. The supply table shows, product by industry, where the supply of each product comes from: domestic output by each industry plus imports, with adjustments for taxes on products and trade and transport margins to reach purchasers' prices. Its transpose, with industries in rows and products in columns, is often called the make table. The use table shows, product by industry, how each product is used: as intermediate consumption by each industry or in final demand.
Because every product's supply must equal its use, the two tables act as a framework for balancing data from different surveys and sources, and the balanced tables underpin annual GDP benchmarks. Statisticians then convert them, using assumptions about technology, into symmetric input-output tables that are either product by product or industry by industry. These symmetric tables are the ones used for the modelling described below.
Final Demand
Final demand is spending on goods and services for final use rather than as inputs into further production. The SNA calls these final uses. Final demand is made up of household and government final consumption, gross fixed capital formation, changes in inventories and exports. Intermediate demand, by contrast, is the purchase of inputs such as steel bought by a car maker or electricity bought by a factory.
The distinction is what prevents double counting in GDP: the steel in a car is counted once, inside the price of the car sold to a household, and not again as a separate purchase. In input-output models final demand is treated as the outside driver of the economy. A change in household spending, in public investment or in foreign demand for exports is the shock whose effects on output in every industry the model traces.
Technical coefficients and the Leontief inverse
Input-output analysis assumes that each industry uses inputs in fixed proportions. The technical coefficient is the value of input from industry needed per unit of output of industry :
Here is the flow from industry to industry and is the total output of industry . Collecting the coefficients in a matrix , the output vector and the final demand vector , the row identity becomes . Solving for output gives:
The matrix is the Leontief inverse. Each column sum is an output multiplier: the total output across all industries needed to deliver one extra unit of final demand for that industry's product, counting direct and indirect inputs.
A two-industry example shows the method. Suppose agriculture uses 0.2 of its own output and 0.4 of manufacturing output per unit, and manufacturing uses 0.3 of agricultural output and 0.1 of its own per unit, so , , and . The determinant of is , and the Leontief inverse has entries 1.5 and 0.5 in the first row and about 0.667 and 1.333 in the second. With final demand of 100 for agriculture and 50 for manufacturing, output must be in agriculture and in manufacturing. As a check, agriculture's output equals . The output multiplier for agriculture is .
Uses and limitations
Input-output models are widely used for economic impact studies of new plants, events or public projects; for estimating how a shock in one sector, such as an energy price rise, spreads through costs; for regional and trade analysis; and, combined with emissions data, for measuring the carbon content of production and consumption. Extended versions feed into satellite accounts and social accounting matrices and into computable general equilibrium models.
The method has clear limits:
- Fixed coefficients. The model assumes no substitution between inputs and constant returns to scale, so it ignores how firms respond to relative price changes.
- No supply constraints. It assumes extra output can always be produced, which overstates effects when the economy is near capacity, and it can exaggerate local multiplier effects if leakages to imports are understated.
- Out-of-date tables. Benchmark tables take years to compile, so the structure they describe may have changed.
Despite these limits, input-output tables remain a basic tool of economic modelling and a core part of the statistical system behind gross value added and GDP.
External links
Ref 1100001802