National Patterns of R&D Resources: 2019–20 Data Update

National Patterns of R&D Resources provides current data on the levels and key trends of the performance and funding of research and experimental development in the United States, with comparisons to the historical record (back to 1953). Detail is also provided for the U.S. total and by performer according to the types of R&D conducted (i.e., basic research, applied research, and experimental development).

Report frequency. Reports in this series are typically published annually.

Current reference period. Data from 1953 to 2020 are provided for almost all variables. The data for 2020 are estimates and not previously reported in this series. The data for 2019 are revised from the previous edition (2018–19) of this series. The data for 1953–2018 reflect final survey data but may still include some further (typically minor) revisions.

The National Patterns statistics draw primarily from the National Center for Science and Engineering Statistics’ (NCSES’s) annual national surveys of the R&D expenditures and funding of the organizations that perform the bulk of U.S. R&D. These organizations include the business sector, federal and nonfederal government, higher education, and nonprofit organizations. Each of these sectors are summarized—including links to other sections of the NCSES website with specifics on survey coverage, variables collected, and survey or sampling implementation—in the sections below.

For 2019, annual data on the R&D performed in the domestic United States by the business sector come from NCSES’s Business Enterprise Research and Development Survey (BERD), covering companies with 10 or more employees, and the Annual Business Survey (ABS), covering companies with 9 or fewer employees. BERD and ABS are conducted for NCSES by the Census Bureau in accordance with an agreement between the two agencies. Both are sample surveys designed to be nationally representative of all for-profit companies, publicly or privately held, in manufacturing and nonmanufacturing industries (https://www.nsf.gov/statistics/srvyberd/, https://www.nsf.gov/statistics/srvyabs/).

For 2017 and 2018, the source of the data for companies with 10 or more employees was the Business R&D Survey (BRDS), a predecessor to the BERD Survey, with similar survey characteristics and coverage. ABS was again the source of data on companies with 9 or few employees in both these years.

From 2008 to 2016, the primary source data on business R&D was the NCSES Business R&D and Innovation Survey (BRDIS), a predecessor of both BERD and BRDS. In a salient difference, BRDIS was designed to be nationally representative of all for-profit companies with 5 or employees. In 2016, a companion survey, the Business R&D and Innovation Survey–Microbusiness (BRDI-M), provided statistics on the R&D activities that year of businesses with 1–4 employees (https://www.nsf.gov/statistics/srvymicrobus/). Prior to the 2016 data year, statistics on the R&D of these “micro” businesses ($4–$5 billion annually) were neither available nor reported in the business R&D totals reported by National Patterns.

As noted, BERD, BRDS, and BRDIS are related surveys. With regard to R&D, the topical coverage has been similar: R&D performance (in the United States and worldwide); total and R&D employment (in the United States and worldwide); sources of R&D funding; type of R&D activities (basic research, applied research, development); type of R&D costs; R&D capital expenditures; R&D application and technology focus areas; geographic location (within the United States and in foreign countries); sales (in the United States and worldwide); and patenting, licensing, other technology transfer activities. The ABS and BRDI-M surveys provided R&D performance and funding data compatible with those for the larger companies.

For 2007 and earlier years, the data come from the NCSES Survey of Industrial Research and Development (SIRD), the predecessor to BRDIS (https://www.nsf.gov/statistics/srvyindustry/sird.cfm). SIRD was also a sample survey, conducted annually (1953–2007), that provided national estimates of the R&D performed within the United States by industrial firms, whether U.S. or foreign owned. The SIRD target population consisted (same as BRDIS) of all for-profit companies with 5 or more employees, manufacturing and nonmanufacturing, that performed R&D in the United States. Data on the R&D of companies with fewer than 5 employees were not available throughout this period of time.

(Note: Data on business intramural R&D capital, as reported by BRDIS and its follow-on surveys, are not included as a component in the National Patterns total of business intramural R&D.)

Federal intramural R&D. Data on the intramural R&D performed by the federal government come from NCSES’s Survey of Federal Funds for Research and Development (Federal Funds Survey), which is completed annually by all federal agencies conducting R&D programs (approximately 28 federal departments or independent agencies; https://www.nsf.gov/statistics/srvyfedfunds/). The Federal Funds Survey data provide intramural R&D totals as outlays and obligations on a federal fiscal year basis. The obligations data are further detailed by federal agency, performer, type of R&D, geographical area, and field of science or engineering (for research, not for development). (Note: data on federal intramural R&D plant reported by the Federal Funds Survey are not included as a component in the National Patterns total of federal intramural R&D.)

Federally funded R&D centers (FFRDCs). The nation’s FFRDCs are a second venue of federal R&D performance. (Currently, there are 42 FFRDCs, although the number can change from year to year; NCSES maintains a Master Government List of the population of FFRDCs, which it regularly updates: https://www.nsf.gov/statistics/ffrdclist/.) FFRDCs are R&D-performing organizations administered by an industrial firm, a university, a nonprofit institution, or a consortium but that have funding exclusively or substantially from the federal government. An FFRDC is operated to provide R&D capability to serve an agency’s mission objectives or, in some cases, to provide major facilities at universities for research and training purposes. Since FY 2001, NCSES’s FFRDC Research and Development Survey (known previously as the Survey of Research and Development Expenditures at FFRDCs) has provided annual data (by federal fiscal year) on the R&D expenditures (with additional detail) of all the nation’s FFRDCs (https://www.nsf.gov/statistics/srvyffrdc/). Prior to FY 2001, R&D expenditure data for the FFRDCs were collected as part of the major performer surveys: the Survey of R&D Expenditures at Universities and Colleges (for university-administered FFRDCs), the Survey of Industrial R&D (for industry-administered FFRDCs), and the Federal Funds Survey (for nonprofit-administered FFRDCs).

The category of nonfederal government R&D performance included in National Patterns is the intramural R&D of state governments (i.e., the R&D performance of state agency and department employees and the services performed by others in support of internal R&D projects). Data on this state intramural R&D come from NCSES’s Survey of State Government Research and Development (https://www.nsf.gov/statistics/srvystaterd/). This state survey is a comparatively recent NCSES initiative, with the first data year in FY 2006.

In addition, the BERD, FFRDC, and Higher Education Research and Development (HERD) surveys (see above and below) provide detail on R&D funding by nonfederal governments other than just state governments.

For academic FY 2010 and onward, data on the R&D performed in higher education come from NCSES’s HERD Survey. HERD is an annual census of universities and colleges that grant the bachelor’s degree or higher and expend at least $150,000 in separately accounted for R&D over the fiscal year of these academic institutions (https://www.nsf.gov/statistics/srvyherd/). The survey collects information on R&D expenditures by field of research, source of funds, type of R&D, and head counts of R&D personnel.

The HERD Survey replaced the earlier Survey of Research and Development Expenditures at Universities and Colleges (https://www.nsf.gov/statistics/srvyrdexpenditures/). For data years FY 2009 and earlier, the Universities and Colleges Survey provides the data on academic R&D.

This sector refers to R&D performed in the United States by nonprofit organizations other than government or academia. Data on the R&D performed by such nonprofit organizations that are federally funded comes from the aforementioned NCSES Federal Funds Survey (reported annually, on a federal fiscal year basis). Data for the R&D performed by nonprofit organizations with funding from within the nonprofit sector and from business sources (both reported on a calendar year basis) are estimated, based on parameters from NCSES’s past comprehensive R&D surveys of the sector. NCSES’s most recent data on nonprofit organization R&D comes from its FY 2016 Nonprofit Research Activities Survey (https://www.nsf.gov/statistics/2021/ncses21202). Before that were the Survey of Research and Development Funding and Performance by Nonprofit Organizations in 1973 and 1996–97 (https://www.nsf.gov/statistics/srvyrdnonprofit/). Data for years prior to 1973 come from earlier National Patterns reports (which reflect, in part, surveys in 1957 and 1960).

Some of the trend analyses in National Patterns draw on National Income and Product Accounts data (e.g., U.S. gross domestic product [GDP], state domestic product) assembled by the U.S. Bureau of Economic Activity (BEA). For details on this U.S. economic data, see http://www.bea.gov/data/gdp.

In keeping with international conventions, U.S. R&D expenditures in current dollars are adjusted for inflation based on BEA’s implicit GDP price deflator.

The National Patterns statistical picture of the U.S. R&D system arises from integration of the primary data from the NCSES sectoral R&D surveys. Adjustments to the primary data are needed in some cases to enable consistent addition. Furthermore, preliminary or otherwise estimated values may be used (and later revised) where final data from one or more of the surveys are not yet available but can reasonably be calculated.

Key features of the methodology include the following.

• Data from the NCSES business R&D surveys (BERD, BRDS, BRDIS, ABS, BRDI-M, SIRD) have been reported on a calendar year basis and are incorporated directly in the National Patterns totals. Data from the Federal Funds Survey and the FFRDC Research and Development Survey are reported on a federal fiscal year basis (1 October–30 September the next year) and are adjusted to the calendar year for National Patterns integration. The data from the HERD Survey are reported on an academic fiscal year basis (typically, 1 July–30 June the next year) and converted to calendar year. Data from the Survey of State Government R&D are reported according to each state’s fiscal year and are also converted to calendar year (presuming that 1 July–30 June the next year is typical).
• NCSES annual surveys of business R&D (particularly BRDIS and SIRD) had long covered the population of private sector companies with 5 or more employees. Inclusion of the BRDI-M data in the National Patterns data on business R&D totals in 2016 expanded the covered population to include companies with 1–4 employees, thereby adding $4.8 billion of R&D to the total reported by BRDIS. BRDS data in 2017 and 2018 covered the population of companies with 10 or more employees. Inclusion of the ABS data for 2017 and 2018 added coverage of companies with 1–9 employees to the National Patterns reported business R&D totals in these years ($5.7 billion and $4.5 billion, respectively) and a similar amount in 2019 ($5.2 billion).
• The method for estimating the type of R&D (i.e., basic research, applied research, experimental development) in the business sector was revised for 1998 and later years. This change resulted in a net decrease in the proportion of business R&D classified as basic research. Accordingly, the data for 1998 and later years are not directly comparable with data for 1997 and earlier years. The transition to BRDIS (data series starting in 2008) does not appear to have introduced further discontinuities in the type-of-R&D estimates. For universities and colleges, the type-of-R&D estimation method was revised in 1998 and again in 2010. Hence, the latest data for higher education are also not directly comparable to those reported in earlier years.
• The data for higher education in 2003 and later years include both science and engineering (S&E) and non-S&E fields. Prior to 2003, only S&E fields are included. Non-S&E R&D was $1.4 billion in FY 2003 and $4.9 billion in FY 2019. Similarly, for the business sector, social science R&D was not included until the 2008 data year, when it was $0.8 billion of the $290.7 billion business R&D total.
• The data for higher education R&D have included, since 1998, a net-out adjustment for R&D funds reported by academic institutions as passed through to other academic subrecipients. This adjustment was recently expanded (back to 1998) to also include pass-throughs to all noneducational recipients (i.e., businesses, nonprofit organizations, and others). The effect of this expanded adjustment is that the annual totals of higher education R&D performance are about $0.4 billion lower in 1998, the first year for this adjustment, to about $2.5 billion lower annually in 2012–15 than previously have been reported in National Patterns. In addition, for the 2010–20 survey years, this more comprehensive pass-through funding adjustment has been further improved by tabulating the adjustments on an institution-by-institution basis, which allows for a more precise estimate of the original nonfederal sources of pass-through funds and better estimation of the type of R&D (basic, applied, or experimental development) being passed through to other R&D performers.
• Data on federally funded R&D discussed in this report were derived from surveys of organizations that perform R&D, such as companies, universities, and FFRDCs. In the past, these amounts differed substantially from the R&D that federal agencies had reported funding. For example, in 2009, federal agencies reported obligating $141 billion for R&D funding to all R&D performers (including $53 billion to the business sector), compared with an estimated $127 billion in federal funding reported by all performers of R&D that year ($40 billion by businesses). More recently, with improvements in the methods for the survey of federal funding and adoption of revised R&D definitions, this effect has become less apparent.
• From 2016 forward, the data on federal intramural R&D explicitly exclude expenditures for pre-production development. (Pre-production development is nonexperimental work on a product or system before it goes into full production, e.g., activities and programs that are categorized as operational system development in the Department of Defense’s research, development, test and evaluation Budget Activity structure.) This change aligns the federal intramural data with a recent change introduced in the definition of R&D by the U.S. Office of Management and Budget. In FYs 2016–18, pre-production development for federal intramural R&D totaled between $5 billion and $7 billion each year.

Additional details on methodology and technical issues pertaining to specific variables are provided in the notes on each table. For further information about compiling the National Patterns statistics, contact the Project Officer.

R&D performers and funders. The U.S. R&D system consists of the R&D activities of differing performers and sources of funding for these activities. The main categories of R&D performers tracked by NCSES are businesses; federal agencies; FFRDCs (administered by businesses, universities, or nonprofit organizations); nonfederal government agencies (specifically those of the 50 states and the District of Columbia); higher education; and nonprofit organizations. For R&D funding, the main categories are businesses; the federal government; nonfederal government (state, regional, local); higher education; and nonprofit organizations. Organizations that perform R&D often receive significant levels of outside funding; R&D funders may also be significant performers.

Type of R&D. As defined by the Organisation for Economic Co-operation and Development’s Frascati Manual (in seven editions since the early 1960s), R&D spans three main types of activities:

Basic research: Experimental or theoretical work undertaken primarily to acquire new knowledge of the underlying foundations of phenomena and observable facts, without any particular application or use in view.

Applied research: Original investigation undertaken in order to acquire new knowledge. It is, however, directed primarily towards a specific, practical aim or objective.

Experimental development: Systematic work, drawing on knowledge gained from research and practical experience and producing additional knowledge, which is directed to producing new products or processes or improving existing products or processes.^{​Organisation for Economic Co-operation and Development (OECD). 2015. Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development. Page 45. Paris: OECD Publishing. Available at http://www.oecd.org/publications/frascati-manual-2015-9789264239012-en.htm.}