USEEIO Emission Factors

Citation-grade reference for the US Environmentally-Extended Input-Output model (USEEIO) and the EPA Supply Chain GHG Emission Factors derived from it. Documents the current production stack as of May 2026: USEEIO v2.5 / v2.5.1 (EPA, April 2025, BEA 2017 benchmark, data year 2022) as the underlying model, EPA Supply Chain GHG Emission Factors v1.3 (July 2024, USD 2022, AR5 GWP-100, 1,016 NAICS-6 commodities) as the most recent EPA-published practitioner dataset, and the Cornerstone v1.4 community-stewarded release (October 2025, Stanford SSL + Watershed) that succeeded EPA stewardship in late 2025. Defines the model’s three-layer architecture, the producer–purchaser price problem, the aggregation bias problem, and the double-counting hazard when EEIO and supplier-specific data are mixed. Aligned to the GHG Protocol Scope 3 Standard for purchased goods and services (Cat 1) and capital goods (Cat 2). Two worked examples — a mid-market professional-services firm computing a full Cat 1 inventory, and the same firm with a single supplier-specific override — show the canonical method and the canonical failure mode. Sector intensity reference table and the fifteen highest-intensity NAICS-6 commodities visualised. Audit checklist for ISO 14064-1 reasonable assurance.

This page is the deep treatment of the USEEIO model and the EPA-derived Supply Chain GHG Emission Factors specifically. The Scope 3 Category 1 spend-based methodology covers the full Cat 1 inventory workflow including supplier-specific data preference order, hybrid approaches, and uncertainty quantification; this page covers the USEEIO factor system itself — how the model is built, how the practitioner factors are derived from it, what the factors do and do not include, and how to apply them without triggering the three named hazards (producer–purchaser price, aggregation bias, double-counting). For the operational tool, see the upcoming Scope 3 Cat 1 Spend-Based Calculator at /calculators/.

Why USEEIO Matters — and Why Two Versions Now Coexist

What Is USEEIO

USEEIO is the U.S. Environmentally-Extended Input-Output model — a family of combined economic–environmental models maintained by the United States Environmental Protection Agency (EPA) Office of Research and Development. The model couples the input-output tables of the U.S. economy (published by the Bureau of Economic Analysis) with environmental satellite accounts (national totals of greenhouse gases, criteria air pollutants, water use, land use, toxics, and material extraction by industry) to produce per-dollar environmental intensity factors for every category of goods and services in the U.S. economy. For GHG accounting practitioners, the most important output of the USEEIO family is the Supply Chain GHG Emission Factors dataset — a per-NAICS-6 table of kg CO₂e emitted per dollar of purchaser spend on each of 1,016 U.S. commodities.

USEEIO is the United States analogue of three other widely-used global EEIO models: EXIOBASE (developed by a European consortium led by NTNU and 2.-0 LCA), GTAP (the Global Trade Analysis Project at Purdue University), and CEDA (the Comprehensive Environmental Data Archive, maintained by Watershed Technology). Each model covers a different geographic footprint and uses different underlying economic and environmental data. USEEIO is the most appropriate model for organisations whose procurement spend originates in the United States.

The model is governed by EPA’s Sustainable Materials Management program and documented in two peer-reviewed publications: the original 2017 Journal of Cleaner Production paper (Yang et al.) and the 2022 Scientific Data paper (Ingwersen et al.) that accompanies the v2.0 release. The codebase is open-source on GitHub (github.com/USEPA/USEEIO), and the practitioner-facing Supply Chain GHG Emission Factors are released as CSV downloads via data.gov.

USEEIO does not measure emissions — it allocates them. The U.S. national GHG inventory (published annually by EPA in the Inventory of U.S. Greenhouse Gas Emissions and Sinks) provides the total emissions by industry that USEEIO then propagates through the supply chain using the BEA input-output tables. The model’s accuracy ceiling is therefore inherited from the national inventory it draws on. What USEEIO adds is the cross-industry attribution: given that an industry emitted X tonnes directly, how much of those emissions are embodied in each downstream commodity by the time it reaches a final purchaser. That allocation is the analytical core of every EEIO model.

Model Architecture — Three Layers

USEEIO’s architecture is best read as three coupled layers. Understanding the layers — and knowing what each layer contributes to the final emission factor — is the prerequisite for diagnosing any methodology question that arises in practice.

BEA Sector Classification and NAICS Mapping

BEA classifies the U.S. economy at multiple resolution levels. USEEIO v2.5 uses two: the Summary level (71 industries, used for the broad architecture and for the annually-updated economic data) and the Detail level (411 industries in the 2017 benchmark, used for the high-resolution v2.5 model variants). The practitioner factors in v1.3 are published at the NAICS-6 level (1,016 commodities), produced by mapping the 411 BEA Detail sectors down to NAICS via the standard BEA-to-NAICS crosswalk. Several BEA Detail sectors map to multiple NAICS-6 codes; in those cases, every NAICS-6 in the group carries the same factor — a known limitation reporters should recognise (see §8 hazard 2).

Three NAICS-6 categories are excluded from the v1.3 factors and from most EEIO factor publications generally: electricity (handled separately via grid factors — see the Scope 2 Electricity methodology), government services (no meaningful “purchaser price” basis), and households (consumption is the model’s final-demand target, not an intermediate input).

LCIA Bridging — From Dollars to kg CO₂e

The economic layer of USEEIO produces, for each commodity, a vector showing how many dollars of each industry’s output are required to produce one dollar of that commodity’s output, traced all the way back through the supply chain (the Leontief inverse, also called the Total Requirements matrix). The environmental layer attaches to each industry a vector of physical emissions per dollar of its output. Multiplying the two produces a per-commodity vector of grams of each GHG embodied in one dollar of that commodity. The LCIA layer collapses that GHG vector to a single CO₂e number by applying GWP-100 characterization factors.

Where B is the environmental satellite matrix (kg of each GHG per dollar of industry output), L is the Leontief inverse / Total Requirements matrix from the BEA input-output tables, y is the final-demand vector isolating the commodity, and GWP_AR5 is the AR5 GWP-100 characterization vector. This is the canonical EEIO equation as set out in Appendix 1 of the EPA Supply Chain GHG Emission Factors report.

The published v1.3 factors split this single number into three components for transparency: the Supply Chain Emission Factor (SEF) — emissions embodied in the production of the commodity itself; the Margins Emission Factor (MEF) — emissions embodied in the wholesale, retail, and transportation margins that get added between producer price and purchaser price; and the SEF + MEF (with margins) total, which is what practitioners apply to most procurement spend. The distinction matters when handling transportation separately (see §9 calculation method).

USEEIO Version Timeline — What Reporters Need to Know

The USEEIO family has evolved through eight identifiable release milestones since first publication in 2017. Reporters with multi-year Cat 1 trend baselines must understand which milestones triggered factor-value drift and which were purely model-improvement releases that left practitioner factors unchanged. Table 5.1 is the standards-document log.

Two structural points for trend analysis. First, any spend-based Cat 1 trend baseline that spans the v1.2 → v1.3 transition (April 2023 → July 2024) crosses a GWP-basis discontinuity. AR4 GWP-100 for methane was 25; AR5 GWP-100 is 28 (a 12% increase). N₂O moved the other direction: AR4 was 298, AR5 is 265 (an 11% decrease). The net effect on any given commodity factor depends on the gas composition of its supply chain — fossil-fuel-intensive supply chains saw modest factor increases, agricultural supply chains saw modest factor decreases. Comparable trend lines require restatement of the baseline year to AR5 basis. Second, the v1.3 → Cornerstone v1.4 transition (July 2024 → October 2025) is a stewardship change, not a methodology overhaul — the analytical engine remains USEEIO v2.5 — but downstream platforms switching libraries in late 2025 introduced a discontinuity at the platform layer that reporters should document in their inventory methodology statement regardless of which library is technically more current.

Method Selection — When to Use USEEIO vs Alternatives

Two structural points. First, USEEIO is rarely the only method on a real inventory — most production Cat 1 disclosures use USEEIO on the long tail of small-value, hard-to-trace spend, while reserving supplier-specific or hybrid methods for the top 20–30 vendors by spend (which typically cover 60–80% of total Cat 1 emissions). Second, “is the supplier U.S.-domiciled” is the wrong test in isolation — what matters is whether the supply chain that produces what you bought looks more like the U.S. economy or another economy. A U.S. distributor reselling imported electronics has a supply chain that looks more like CEDA’s coupled-imports treatment than USEEIO domestic-only. The v2.5 “kingbird/catbird” coupled-imports model variants exist for exactly this reason; reporters using the most recent factor sets benefit from this without needing to construct it themselves.

USEEIO in the GHG Protocol Scope 3 Categories

The GHG Protocol Corporate Value Chain (Scope 3) Standard Chapter 7 names spend-based EEIO calculation as the screening-grade method for two upstream categories and recognises it as defensible for several others. Table 7.1 maps USEEIO applicability across the fifteen Scope 3 categories.

Limitations and Known Hazards — Read Before Reaching for Factors

Three hazards account for the great majority of audit findings on USEEIO-based Cat 1 disclosures. Each is named for citation and has a structural detection signature.

The Calculation Method — Step by Step

The full method for computing a Cat 1 line using USEEIO factors has four steps. Each step has a discrete audit deliverable.

Step 1 — Identify procurement spend and assign NAICS-6 code

Pull procurement spend from the accounting system, ideally categorised by the supplier’s primary NAICS code on the vendor master record. Where NAICS coding is absent, map procurement categories or general ledger lines to the closest NAICS-6 description. The factor’s resolution is NAICS-6 — finer categorical splits do not improve accuracy; coarser ones lose specificity.

Step 2 — Apply purchaser-to-producer price adjustment (currency-year deflation)

Deflate the spend to the factor’s currency year (USD 2022 for v1.3). Use BLS CPI-U All Items or BEA GDP deflator. Document the index values. For 2024 USD spend applied to v1.3 factors: multiply by 292.655 / 313.689 ≈ 0.9329. This is the producer–purchaser price problem prevention step (hazard 8.1).

Step 3 — Apply the emission intensity factor

Multiply deflated spend by the SEF + MEF (with margins) factor for the matching NAICS-6 code. Use the “with margins” variant unless transportation has been handled separately, in which case use SEF (without margins). The factor unit is kg CO₂e per 2022 USD purchaser price; the result is kg CO₂e.

Step 4 — Aggregate and classify by Scope 3 category

Sum across all spend lines. Classify each as Cat 1 (operating spend on purchased goods and services) or Cat 2 (capital goods). The factors themselves don’t differ between Cat 1 and Cat 2 — what differs is the inventory line the result is reported under.

Worked Examples — Audit Records

Two examples at this page’s review date (2026-05-16), both for a fictional mid-market U.S. professional-services firm with FY2024 procurement. Hardcoded values throughout — these are audit-record snapshots and must reconcile to stated inputs regardless of future factor releases. Example 10.1 walks the canonical full-EEIO Cat 1 inventory. Example 10.2 introduces a single supplier-specific PCF and walks both the correct integration and the double-counting trap (hazard 8.3).

USEEIO Emission Intensity Factors — BEA Summary-Level Reference

Table 11.1 lists the BEA Summary-level emission intensities for the 24 most-cited sectors at the practitioner level — the categories that appear most often on procurement spend reports. Values are derived from EPA Supply Chain GHG Emission Factors v1.3 (kg CO₂e per 2022 USD, purchaser price, with margins, AR5 GWP-100). For a specific NAICS-6 commodity within a BEA Summary group, consult the raw EPA CSV (linked at the bottom of the table). Where a Summary group contains highly heterogeneous sub-sectors, both a typical value and a range are shown.

Need a specific NAICS-6 commodity factor not shown above? Download the full EPA Supply Chain GHG Emission Factors v1.3 CSV (data.gov) — 1,016 commodities with separate SEF, MEF, and SEF+MEF columns. The raw dataset is the citation-grade source.

The fifteen highest-intensity NAICS-6 commodities — visual

USEEIO vs Other EEIO Databases — Comparison

Practitioners outside the U.S. — or with global supply chains — sometimes need EEIO factors from other published models. Table 12.1 is the canonical cross-database reference. Models are not interchangeable: each is calibrated to a different economy, with different sector classifications and different GHG vintages.

The selection rule is geographic-and-purpose-driven, not quality-driven. USEEIO is the right model for the U.S. supply-chain footprint of a U.S.-domiciled reporter; EXIOBASE is the right model for the EU footprint of an EU reporter; CEDA is the right model for a global reporter with material supply chains in emerging markets. Mixing models within one inventory is acceptable but requires explicit disclosure: most major reporters with diversified operations apply USEEIO to their U.S. spend, EXIOBASE to their EU spend, and CEDA or USEEIO v2.5-coupled-imports to the residual. The methodology statement names the mix.

Standards Alignment

Five standards govern the methodological status of USEEIO-derived Cat 1 disclosures.

What the Calculator Handles vs What You Decide

The Scope 3 Cat 1 Spend-Based Calculator automates the mechanical steps. This methodology page covers the upstream decisions the calculator cannot make for you.

Audit Checklist — USEEIO-Based Reasonable-Assurance Pre-Flight

Eight items, USEEIO-specific. Run before submitting any Cat 1 / Cat 2 disclosure for reasonable-assurance review under ISO 14064-1 or equivalent.

Methodology Metadata — for GHG Inventory Documentation

Copy verbatim into your GHG inventory methodology statement for ISO 14064-1 transparency compliance. Adjust the version, deflator, and basis lines to match your inventory choices.

Frequently Asked Questions

Calculator

Apply EPA v1.3 / Cornerstone v1.4 factors with automatic NAICS-6 lookup, CPI deflation, SEF/MEF selection, hybrid supplier-specific override handling, and audit-grade output: Scope 3 Cat 1 Spend-Based Calculator — currently in development at /calculators/. In the meantime, the methodology above can be applied in any spreadsheet; the worked example shows the full calculation chain.