Input-Output Analysis
- Introduced by
- Wassily Leontief
- Year
- 1936
- Type
- Linear inter-industry accounting and impact model
- Family
- Inter-industry analysis
What does it do?
Input-output analysis is a quantitative framework for representing the interdependence between the industries of an economy, introduced by Wassily Leontief in 1936.
It records the flows of goods and services between sectors in a transactions table, derives fixed technical coefficients describing how much each industry buys from every other industry per unit of output, and inverts the resulting linear system to trace how an exogenous change in final demand ripples through the entire production structure.
The intuition behind the method
Imagine the economy as a web in which producing one extra car requires steel, the steel requires coal and electricity, the electricity requires more steel, and so on without end. Input-output analysis cuts through this infinite regress with a single matrix inversion: the Leontief inverse sums up every round of indirect requirements at once, so you can ask 'if households want one billion dollars more cars, how much total output must every sector produce?' and get an exact answer that includes the direct demand plus all the supply-chain echoes it triggers.
History
Wassily Leontief, a Russian-American economist at Harvard, published the foundational input-output framework in 1936 in the Review of Economics and Statistics, building the first empirical inter-industry table for the 1919 U.S. economy. The work, which operationalized Walras's abstract general-equilibrium vision into something estimable from data, earned Leontief the 1973 Nobel Memorial Prize in Economic Sciences. Input-output accounts were institutionalized in national statistical systems after World War II and remain a core component of the System of National Accounts worldwide, later generalized into social accounting matrices and embedded inside computable general equilibrium models.
When to use it
Use input-output analysis when you need to quantify the economy-wide ripple effects of a change in final demand, a new investment, a policy shock, or a supply disruption, and you have access to an inter-industry transactions table (often published by national statistical offices every few years). It is the standard tool for impact assessment of large projects, regional economic studies, environmental and carbon footprint accounting, and supply-chain analysis. It assumes fixed input proportions, no capacity constraints, no price responses, and constant returns to scale, so it is best for short-run, demand-driven impact questions rather than long-run analyses where substitution and prices matter — for those, a computable general equilibrium model is more appropriate.
Limitations and considerations
- Assumes fixed technical coefficients and linear, constant-returns technology, ruling out input substitution and economies of scale.
- Demand-driven and price-free: it cannot model supply constraints, capacity limits, or how prices reallocate resources.
- Coefficients are fixed at the table's base year and become outdated as technology and trade patterns change.
- Tables are costly to compile and are published infrequently and at a coarse sectoral aggregation, masking heterogeneity within sectors.
Source conditions do not replace independent scientific review or validation by an analysis engine.
Strengths
- Captures the full chain of direct and indirect inter-industry effects in a single, exact matrix inversion, avoiding double counting.
- Built directly on official national accounts, giving it an empirical, internally consistent accounting foundation.
- Transparent and reproducible: multipliers can be decomposed and audited sector by sector.
- Readily extended to satellite accounts for employment, energy, emissions, and water, making it a workhorse for environmental and regional analysis.
Applications
- Estimating the total output, employment, and income effects of a major infrastructure project or industry expansion through impact multipliers.
- Constructing environmentally extended input-output models for carbon footprints, energy use, and emissions embodied in trade.
- Regional and inter-regional analysis of economic linkages and the local impact of plant openings or closures.
- Tracing the propagation of supply-chain disruptions and disaster impacts across interconnected industries.
- Providing the structural backbone for computable general equilibrium and social accounting matrix models used in policy simulation.
How it works
How it works
The accounting backbone is that each sector's total output x_i equals what it sells as intermediate inputs to all other sectors (the sum of inter-industry sales z_ij) plus what it sells to final demand f_i (households, government, investment, exports). This balance must hold for every row of the transactions table.
Formula
x_i = \sum_{j} z_{ij} + f_iDividing each inter-industry flow z_ij by the total output x_j of the purchasing sector yields the technical coefficient a_ij — the value of input from sector i needed to produce one unit of output in sector j. Collected into the matrix A, these coefficients are assumed fixed (a Leontief fixed-proportions production technology), which linearizes the system.
Formula
a_{ij} = \frac{z_{ij}}{x_j}, \qquad A = [a_{ij}]Substituting the coefficients into the balance identity gives x = Ax + f: total output equals intermediate use generated by output itself plus final demand. Rearranging isolates final demand as (I − A)x = f, the central Leontief equation linking the production vector to the demand it must satisfy.
Formula
x = A x + f \;\Longleftrightarrow\; (I - A)\,x = f
Solving for output requires inverting the matrix (I − A). The resulting Leontief inverse L = (I − A)^{-1}, also called the total-requirements matrix, has entries ℓ_ij giving the total (direct plus indirect) output of sector i needed to deliver one unit of sector j to final demand. The power series I + A + A² + … shows it accumulates every successive supply-chain round.
Formula
x = (I - A)^{-1} f, \qquad L = (I - A)^{-1}Summing each column of the Leontief inverse produces the output multiplier m_j: the total economy-wide output generated per unit of final demand for sector j. Analogous Type I and Type II multipliers for employment, income, and value added are built by pre-multiplying the inverse with the corresponding direct coefficients.
Formula
m_j = \sum_{i} \ell_{ij}Common pitfalls
- Treating impact multipliers as net economic benefits — they measure gross output generated, not welfare or opportunity-cost-adjusted gains.
- Applying national coefficients to a region without adjusting for cross-border purchases (regional purchase coefficients), which overstates local multipliers.
- Ignoring the import leakages embedded in intermediate use, which inflates domestic impact estimates.
- Extrapolating with a stale base-year matrix when the underlying technology or supply structure has shifted substantially.
Frequently asked questions
What is the difference between input-output analysis and a computable general equilibrium (CGE) model?
Input-output analysis uses fixed technical coefficients and a linear, demand-driven structure with no prices, so output responds proportionally to final demand. A CGE model embeds the same accounting structure but adds optimizing agents, flexible substitution between inputs, market-clearing prices, and resource constraints, allowing it to capture substitution and reallocation effects that the linear I-O model holds fixed. I-O is the calibration starting point; CGE is the behavioral extension.
What does the Leontief inverse actually represent?
Each entry ℓ_ij of the Leontief inverse (I − A)^{-1} is the total output that sector i must produce, directly and indirectly, to deliver one unit of sector j's product to final demand. It equals the infinite sum I + A + A² + A³ + …, where each term is a successive round of inter-industry requirements, so the inverse compactly captures the entire supply-chain cascade.
Why are output multipliers often criticized in impact studies?
Multipliers measure gross output generated per unit of final demand, not net welfare gains, and they assume idle resources, no price changes, and no crowding out. Used uncritically — for example by ignoring import leakages or applying national coefficients to a small region — they systematically overstate the true economic benefit of a project, which is why careful analysts adjust coefficients and report assumptions explicitly.
Related methods
Sources
Leontief, W. W. (1936). Quantitative input and output relations in the economic system of the United States. The Review of Economics and Statistics, 18(3), 105–125.
Open DOI referenceCopied.Miller, R. E., & Blair, P. D. (2009). Input-Output Analysis: Foundations and Extensions (2nd ed.). Cambridge University Press.
ISBN: 9780521739023Copied.