Fine scAle eConomic daTa
FACT
FACT develops reproducible datasets and estimation routines for fine-scale economic structure. It uses detailed and up-to-date official statistical and remote-sensing information to estimate employment, input-output relations, and domestic trade flows at fine spatial and industrial scales. The estimation procedures are based on explicit representations of the economic mechanisms that shape the localization and organization of economic activity.
Approach
Reproducible analysis workflow
FACT workflows are based on official statistical data sources and openly accessible code, documentation, and service containers. The data services are structured so that analyses can be regenerated from source data and updated when new official periods become available. Estimation procedures are specified separately for each application and documented with their data requirements, assumptions, and validation checks.
European employment estimation
The European employment workflow uses official Eurostat employment totals, with U.S. observed fine-scale employment as training evidence at ISIC section detail. Child shares are estimated from mechanism-derived variables representing settlement, infrastructure, accessibility, natural-resource endowments, and inter-industry dependence, using constrained candidate models combined by a held-out convex ensemble. Predictions are applied stepwise from country to NUTS 1, NUTS 2, NUTS 3, and LAU while preserving official parent totals exactly and reporting uncertainty from geographically blocked cross-validation.
U.S. county IO and domestic trade flows
The U.S. county workflow uses BEA make/use tables before redefinitions at producers' prices, with BEA summary- and detail-level products. County production-side values are distributed using employment shares from QCEW, final demand using population shares from the Census, and domestic trade flows are estimated with a balancing cross-hauling trade-flow methodology using county-to-county road travel times. The methodology reference is doi:10.1080/00343404.2017.1286009.
European regional IO and trade flows
The European NUTS 3 workflow uses FIGARO national industry-by-industry tables at ISIC section detail. NUTS 3 production-side values are distributed using the employment estimates described above, final demand using population shares, and continental trade flows are estimated with a balancing cross-hauling trade-flow methodology using NUTS 3-to-NUTS 3 road travel times, with cross-border flows aligned to the nonnegative country-pair totals reported by FIGARO. The methodology reference is doi:10.1080/00343404.2017.1286009.
Codebase
Source code and containers
Repository group
FACT source code, database build files, service containers, and analysis routines are open source and licensed. Repositories are organized under the FACT GitLab group.
Results
Published datasets
US county IO and domestic trade flows, 2021
County-level regional IO tables and domestic trade flows for the United States in 2021, covering 3,143 counties and county-equivalent areas, at BEA summary-level industry detail (67 summary categories).
US county IO and domestic trade flows, 2017
County-level regional IO tables and domestic trade flows for the United States in 2017, covering 3,143 counties and county-equivalent areas, at BEA detail-level industry detail (389 detail categories).
European employment estimates, 2021, country-NUTS1-NUTS2-NUTS3-LAU, ISIC section level
Stepwise European employment estimates for 2021 from country totals to 123 NUTS 1, 330 NUTS 2, 1,501 NUTS 3, and 98,116 LAUs across 35 countries, at FACT ISIC section detail (22 categories). The release includes uncertainty measures and parent-child conservation validation.
European NUTS3 regional IO tables and trade flows, 2021, ISIC section level
Europe-wide regional input-output tables and continental trade flows for 2021, covering 1,499 NUTS 3 regions in 34 countries, at ISIC section detail (19 categories), anchored to FIGARO national tables and checked against FIGARO country-pair trade totals.
Citation
Authorship and DOI
Author
Riccardo Boero. Single components and result datasets have their own DOI. This project was initiated with the generous support of a SIS internal project from NILU.
@misc{boero_fact,
author = {Riccardo Boero},
title = {FACT: Fine scAle eConomic daTa},
year = {2024},
doi = {10.17605/OSF.IO/PV4ZW},
url = {https://doi.org/10.17605/OSF.IO/PV4ZW}
}