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

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