Note: Updates to this data product are discontinued.
Data for public and private funding of food and agricultural research and development cover the years 1970-2015 (public) and 1970-2014 (private). Data are available as nominal figures and as figures adjusted for inflation.
Public funding is based on data from two sources:
- Data on USDA Federal research are from the National Science Foundation's (NSF) Survey of Federal Funds for Research and Development that reports research and development obligations and outlays to USDA's agencies. To provide the most consistent long-term series, these data are obtained from the National Science Foundation's Integrated Science and Engineering Resources Data System, interactive data tool.
- State-level data are from USDA's Current Research Information System (CRIS), with one exception. Funding to the States from the National Institute for Food and Agriculture (NIFA) is now also from NSF’s Survey of Federal Funds for Research and Development. For the years 1970-1992, the non-NIFA State level data come from the "Inventory of Agricultural Research" publications from CRIS. CRIS stopped publishing the Inventory after 1997 and moved to a web-based reporting system that can be found on the CRIS home page. Data for 1993-97 overlap in the Inventory and the web-based system. In these estimates, data from the web-based system are used for 1993-2015.
In the NSF data, research obligations (amounts for orders placed, contracts awarded, services received, and similar transactions during a given period, regardless of when the funds were appropriated) and research outlays (amounts for checks issued and cash payments made during a given period, regardless of when the funds were appropriated) are reported for each department (e.g., USDA) and agency within a department (e.g., the Agricultural Research Service (ARS) or NIFA). In previous versions of the ERS series, public funding was based on reported outlays, as actual expenditures are most likely to result in research impact. However, certain useful distinctions (e.g., intramural vs. extramural research) are available only in the NSF series for obligations. In the ERS estimates, outlays for a given year are assumed to be divided between intramural and extramural expenditures in the same proportions that obligations are divided in that year. In any given year, outlays may be divided in different proportions than obligations, but the expected value of the differences over many years will be zero if there are no systematic biases in the transformation of obligations to outlays.
Federal-level intramural research is estimated by the NSF series for intramural research for all USDA agencies except NIFA, adjusted from obligations to be consistent with department and agency outlays, as described above.
NIFA funding to States has been reported differently by CRIS in recent years. In some years, only block grants ("formula funding") have been reported by CRIS, while competitive grants have not been reported. Therefore, current ERS estimates of State-level expenditures funded by NIFA are estimated from NSF data. However, in the NSF data, some NIFA funding is reported as "intramural." In the ERS data series, estimates of NIFA funding depart from this NSF convention and regard all NIFA funding, including NIFA administrative costs, as going to the States. This is consistent with CRIS reporting conventions. The only NIFA funding not assumed to be received by the States is funding to firms through the Small Business Innovation Research (SBIR) program.
Thus, total State-level performance estimates are the result of adding estimates of research funded from all sources other than NIFA that were derived from the CRIS system to estimates of NIFA funding to States constructed from NSF data as described above.
CRIS reporting requirements changed in FY 2010. Since that year, some States have not been fully reporting agricultural research expenditures from some non-NIFA sources (e.g., other non-USDA Federal sources or money appropriated by State legislatures).
Private funding estimates are constructed by ERS, based on the methodology presented in the ERS report:
- Research Investments and Market Structure in the Food Processing, Agricultural Input, and Biofuel Industries Worldwide (ERR-130, December 2011).
This report built on earlier work contained in Private-Sector Agricultural Research Expenditures in the United States, 1960-92 (Staff Paper AGES9525). Research expenditures are estimated for eight agricultural industries and then aggregated. These industries include seven agricultural input industries: 1) crop genetic improvement; 2) crop protection chemicals; 3) synthetic fertilizers; 4) farm machinery; 5) animal health; 6) animal genetic improvement; and 7) animal nutrition; and the agricultural output industry, 8) food and kindred products.
We used a number of approaches to construct estimates of private R&D spending by sector. For research-intensive agricultural input industries, we built a database of agriculturally-related research spending firm by firm over time, for all firms in the sector (including "legacy" firms, or firms that exited the industry during the period of study) that have or have had significant R&D expenditures. For large conglomerates, for which agriculture may be only one business segment, we separated agriculturally related R&D spending from R&D spending on nonagricultural business segments. We gathered this information by canvassing a broad set of materials, including company annual reports and websites, reports by industry associations and consulting services, and personal interviews with company representatives.
For agricultural input industries in which firms do not often report their research spending, we estimated agricultural R&D for the industry by taking a percentage of total agricultural input sales, with the percentages (or research intensities) derived from observations on R&D spending from a subset of firms and from previous surveys of the industry. Estimates for the food manufacturing industry are based on the U.S. country-level estimate produced by the Organisation for Economic Co-operation and Development and NSF estimates from the Business Research and Development Survey (BRDIS), as well as the earlier Survey of Industrial Research and Development.
Since private agricultural input research is more likely to influence agricultural productivity than private food manufacturing research, data from the seven agricultural input industries are aggregated into one series, and R&D for food manufacturing is presented in another series. However, the NSF and OECD estimates for food manufacturing research include firms performing animal nutrition research in their sampling universe. Thus, if the agricultural input series is added to the food manufacturing research, conceptually it will double count animal nutrition research. To avoid this double counting, the series "total private food and agriculture R&D" is adjusted to count animal nutrition research only once.
The R&D deflator (to convert current dollars (nominal) to real dollars (adjusted for inflation)) is based on methodology described in Pardey et al. (U.S. Agricultural Research Deflators: 1890-1985), Research Policy 18:5 (1989): 289-296, which provides weights for three components: researcher salaries, State and local government consumption expenditures, and construction costs. The deflator has been updated using information from the American Association of University Professors (The Annual Report on the Economic Status of the Profession, various years) and the Bureau of Economic Analysis (National Economic Accounts, Table 3.9.4: Price Indexes for Government Consumption Expenditures and Gross Investment).