Freight Analysis Framework (FAF5) Regionalized Input-Output Accounts for the Contiguous United States, 2017 and 2022

Description

This dataset provides multi-regional input-output accounts for the contiguous United States, regionalized to the 132 Freight Analysis Framework version 5 (FAF5) zones defined by the Bureau of Transportation Statistics. Accounts are provided for 2017 and 2022. The data are intended for environmentally extended input-output analysis and related applications requiring a spatially disaggregated representation of the U.S. economy. Files: Each file is provided separately for 2017 and 2022, denoted by _{year} in the filename. All monetary values are in millions of USD. FAF zone codes are zero-padded three-digit strings matching the FAF5 zone identifier scheme. BEA codes follow the BEA Summary-level commodity and industry classification. A_regionalized_{year}.csv — Technical coefficient matrix A, with dimensions (n_zones × n_commodities) × (n_zones × n_commodities). Row and column indices are a two-level (faf_zone, bea_code) MultiIndex. Element A[i,j] gives the input of commodity i from region r required per unit of output of commodity j in region s. All column sums are less than 1. This is the core input required to construct the Leontief inverse. q_regionalized_{year}.csv — Commodity output vector q. Each row corresponds to a (faf_zone, bea_code) node and gives total commodity output in USD millions for that zone-commodity pair. g_regionalized_{year}.csv — Industry output vector g. Each row corresponds to a (faf_zone, bea_code) node and gives total industry gross output in USD millions for that zone-industry pair. Y_final_demand_{year}.csv — Final demand table in long format with columns bea_code, faf_zone, and final_demand_USD_millions. Each row gives aggregate final demand for one commodity in one FAF zone. a_N_direct_requirements_{year}.csv — The national direct requirements matrix from the BEA Use table, provided as a reference. Rows are commodities, columns are industries. This is the pre-regionalization input-output structure from which the MRIO system is derived. faf_zones_{year}.csv, commodities_{year}.csv, industries_{year}.csv — Dimension lookup tables listing the FAF zone codes, BEA commodity codes, and BEA industry codes. The ordering of entries in these files defines the row and column ordering of all matrices and vectors above. Users reconstructing the MRIO system programmatically should load these files first and use them to align all other objects. Important note on ordering: All matrices and vectors are stored in zone-major order: all commodities for zone 1 appear first, followed by all commodities for zone 2, and so on. The flattened index for any matrix or vector is therefore equivalent to the sequence [(zone, commodity) for zone in faf_zones for commodity in commodities]. Users must preserve this ordering when constructing the Leontief inverse or any environmental extension. Skeleton code for loading these files, constructing the Leontief inverse, and attaching an environmental extension vector is provided in io-import.py

Steps to reproduce

Detailed methods are presented in io-disaggregation-documentation.docx. Quick summary: The national input-output accounts are sourced from the Bureau of Economic Analysis (BEA) Summary-level Make and Use tables for 2017 and 2022: https://apps.bea.gov/iTable/?reqid=1602&step=2&Categories=Core&isURI=1. Regionalization follows a trade-share proportional allocation method in which the BEA national Make and Use tables are downscaled to FAF zones using origin-destination freight flow shares derived from the FAF5 dataset. Regional commodity production is allocated across zones in proportion to each zone's share of observed outbound freight flows; regional intermediate demand is then derived from national input recipes applied to regional output. Origin-destination trade flows are used to construct the full interregional supply chain structure, linking commodity production in each origin zone to intermediate and final demand in each destination zone. Service sectors, which are absent from FAF5, are treated as intraregional and allocated to zones proportional to population. The resulting system preserves national totals from the BEA accounts while capturing the geographic structure of U.S. commodity production and trade. The paper related to this dataset is currently in preparation. You can cite it like this in the meantime: Fowler KF, R. RR, R. BL, and L. C, “Tracking ecological dependencies in trade: Embodied human appropriation of net primary production (eHANPP) for U.S. counties,” In preparation, 2026.

Institutions

• Northern Arizona University

  Arizona, Flagstaff

Categories

Funders

• U.S. National Science Foundation

  Government of the United States of America

  Alexandria

  Grant ID: CBET-2115169

Licence