The Agri-Food Economic Data System (Ag-FEDS), developed by ERS, describes how commodities are made, used, and distributed across the domestic economy. Specifically, Ag-FEDS shows how more than 350 production activities (e.g., grain farming, breakfast cereal manufacturing, and truck transportation) culminate in broad and detailed annual food and beverage expenditures.
Ag-FEDS was developed to support multiplier analysis on the agri-food supply chain, with applications such as the ERS Food Dollar and the Resource Requirements of Food Demand data. The Food Dollar multiplier model computes the cost distribution (or, equivalently, value added) across agri-food supply chain activities that produce, process, and distribute food and beverage products to final markets for consumer purchase.
- Data Sources
- Data Coverage
- Methods
- Strengths and Limitations
- Resources
- Recommended Citation
- File Downloads
Data Sources
Ag-FEDS is compiled with data from the Bureau of Economic Analysis (BEA) input–output tables, as well as from ERS (see “Related Data” in the Food Dollar Documentation). The ERS report, Documentation for the Agri-Food Economic Data System (Ag-FEDS): A More Complete Accounting of the U.S. Agri-Food Economy, provides a complete list of data sources for Ag-FEDS.
- Data published by the Bureau of Economic Analysis
- Make and Use tables, Industry Gross Output table, and Value Added table published in benchmark years beginning with 2007 at the detail level and annually at the summary and summary underlying levels
- Personal Consumption Expenditure Bridge tables published at the 2007, 2012, and 2017 benchmark detail and annual summary levels
- Personal Consumption Expenditures by Product annual tables published in the National Income and Product Accounts
- Margins tables published at the 2007, 2012, and 2017 benchmark detail level
- Data published by ERS
Data Coverage
Ag-FEDS data are national data available annually for the years 2007–24 and contain market value data showing the assembly of 360 commodities by 358 production activities. Commodities are finished or semi-finished goods and services. For example, sweet corn and field corn are two distinct goods. Sweet corn can be purchased fresh, canned, or frozen. Field corn, by contrast, is processed and used as an input into a variety of other goods, such as corn syrup, corn meal, and livestock feed. Field corn is captured in commodity C002, grain farming, and sweet corn is captured in C003, vegetable farming. Activities are groups of establishments that produce commodities using similar production processes. The grain farming activity, A002, includes establishments that produce field corn, and the vegetable farming activity, A003, includes establishments that produce sweet corn. While both activities produce corn, their different production processes and final markets make them different production activities. Ag-FEDS make and use data cover the entire U.S. economy and are consistent with BEA benchmark and annual data in the Industry Economic Accounts. Ag-FEDS data comprise:
- 358 activity accounts and 360 commodity accounts detailing production flows
- 15 voucher accounts that align retail purchases of food with food-related transactions by activities that sell food or food service in addition to their primary outputs
- 45 final demand accounts
- 41 food- and beverage-specific consumer spending categories
- 28 food at home
- 3 alcohol at home
- 7 food away from home
- 3 alcohol away from home
- 4 nonfood final demand categories
- Personal consumption expenditures
- Private investment and expenditures
- Government expenditures
- Exports
- 41 food- and beverage-specific consumer spending categories
- 3 value added accounts
- Salaries and wages
- Output taxes less subsidies
- Operating surplus
- Imported inputs used in domestic production
Methods
Ag-FEDS is assembled from BEA’s Industry Economic Accounts, including benchmark and annual Input–Output Make, Use, Margin Details, Gross Output, and Value Added tables, benchmark and annual Personal Consumer Expenditures (PCE) Bridge tables, and annual PCE-by-product tables from the National Income and Product Accounts (NIPA). These tables report nationally aggregated data for production, final demand, and factor payment transactions at market values. The annual estimates are more broadly defined whereas the benchmark estimates are more narrowly defined and detailed but available on a less frequent basis (i.e., every 5 years with the release of the Economic Census). Using constrained maximum likelihood estimation, Ag-FEDS replicates published annual and benchmark BEA data and imputes data for benchmark accounts on an annual basis.
Ag-FEDS accounts are reaggregated from BEA’s accounts to reflect the national economy with special focus on the food system. Ag-FEDS data are reconciled with each benchmark when new Economic Census data are released. Additionally, each annual data release may revise previous data with revisions in source data. The total possible number of activity and commodity accounts in the published Ag-FEDS data set is 365; in a given year, however, there may be fewer, based on BEA aggregation in Economic Census years. Relative approximation errors are computed for constrained maximum likelihood estimates at the transaction level and provide an assessment of the reliability of the Ag-FEDS data in non-benchmark years. Because Ag-FEDS imputes detailed accounts in non-benchmark years, uncertainty increases with each year past the benchmark. A full description of the estimation procedure, error statistics, and account reaggregation is provided in Rehkamp and Canning (2025).
Ag-FEDS incorporates ERS Price Spreads from Farm to Consumer and Meat Price Spread data to improve the measurement of foodservice value and food and beverage commodity value in production and final markets. These measurement issues can occur in BEA data when food and beverages are prepared and sold as secondary outputs by places such as sports venues, schools, and hospitals. Using ERS price spread data, Ag-FEDS redirects the value of food and beverage sales from secondary foodservice activities to their primary sources of production. Additionally, Ag-FEDS harmonizes BEA PCE estimates with the ERS Food Expenditure Series to match consumer food spending dollar for dollar. This harmonization reconciles BEA accounting practices that tend to overestimate the value of food in food at home and underestimate the value of food in food away from home with survey data from retailers that directly measure food spending (Rehkamp & Canning, 2025; Okrent et al., 2018).
Ag-FEDS supports matrix computation methods for Type I multiplier models. Type I multipliers measure the direct and indirect effects of changes in demand, but they do not measure induced effects from changes in household incomes in response to changes in production. The annual nature and rich level of detail of the Ag-FEDS data means these multipliers can be computed to show changes in production requirements in the domestic economy for changes in demand, as well as subsets of production activities and demand, such as agri-food supply chain responses to changes in food demand (Rehkamp et al., 2026; Miller & Blair, 2022).
Strengths and Limitations
Ag-FEDS is a comprehensive, detailed picture of annual economic activity aggregated to the national level with special attention to the food system. Ag-FEDS provides researchers an opportunity to study changes in the industry structure of domestic food production and the commodity composition of food demand over time. For example, the data system shows value contributions of commodity and factor inputs, such as farm commodities, manufacturing and processing, and labor to the cost of finished food products like cereals and beef. Further, Ag-FEDS data show the activity production requirements for food commodities. For example, Ag-FEDS reports the market value of activity output required to produce breakfast cereals and beef cattle ranching. This is useful for identifying the production and margin costs of individual commodity inputs to food products over time. While Ag-FEDS is focused on the food system, these data are available for the whole economy. Interested users could also measure, for example, these same value and cost contributions to textile production and demand over time. The multipliers generated from Ag-FEDS can be used to show changes in total production requirements for changes in demand and subsets of demand on a per dollar basis. Rehkamp et al. (2026) provides example R code for generating the multipliers for a year of data.
Because Ag-FEDS uses nationally aggregated data from the BEA, it does not report market values at the sub-national geographic level, nor does it report values for specific crops used in food, such as corn and wheat. These are aggregated with other crops in a single farm commodity (i.e., grain farming). The Ag-FEDS aggregation of commodities and activities is dependent on BEA’s aggregation of them by the North American Industry Classification System every 5 years with the release of the Economic Census. The level of detail released by BEA can change for several reasons, such as changes in the Economic Census survey methodology or changes in consumer demand over time. When the level of detail in the source data changes, Ag-FEDS will reflect this change, and some detail may not be available that was previously reported.
Resources
Rehkamp, S., & Canning, P. (2025). Documentation for the Agri-Food Economic Data System (Ag-FEDS): A more complete accounting of the U.S. agri-food economy (Report No. TB-1973). U.S. Department of Agriculture, Economic Research Service.
Rehkamp, S., Zachary, J. C., Baker, Q., & Canning, P. (2026). Documentation for the Agri-Food Economic Data System (Ag-FEDS): Multiplier model applications (Report No. TB-1974). U.S. Department of Agriculture, Economic Research Service.
Okrent, A., Elitzak, H., Park, T., & Rehkamp, S. (2018). Measuring the value of the U.S. food system: Revisions to the Food Expenditure Series (Report No. TB-1948). U.S. Department of Agriculture, Economic Research Service.
Miller, R. E., & Blair, P. D. (2022). Input–output analysis: Foundations and extensions (third edition). Cambridge University Press.
Recommended Citation
U.S. Department of Agriculture, Economic Research Service. (2026). Agri-Food Economic Data System (Ag-FEDS) [data set].
Note that the machine-readable data file provided in csv format should be used with statistical programs capable of processing large data sets.