Introduction State Distribution of R&D Performance Ratio of R&D to Gross State Product Federal Support for R&D Technical Note State and Engineering Profiles
Economists and policy makers have long considered research and development (R&D) to be a key component of economic growth. The contribution of R&D activities to local economies has been a topic of particular interest to state policy makers. This report, Science and Engineering State Profiles: 1997, provides statistics on the geographic distribution of R&D within the United States. R&D data for 52 areas-each of the 50 states, the District of Columbia and Puerto Rico-are derived from the several performer-based surveys of the National Science Foundation's (NSF's) R&D Statistics Program. For each state (or geographic area) these data are categorized by major source of funds (industry, Federal Government, and academia), and by type of performer [industry, Federal Government, academia, Federally Funded Research and Development Centers (FFRDCs), and other nonprofit institutions] (table 1). Data pertaining to Federally-funded R&D are presented in greater detail, in the form of a cross tabulation between: (1) specific Federal agencies from which such funds originate, and (2) the type of performer receiving those funds.
The most recent R&D data available on a state-by-state basis are for the year 1995. In that year, total R&D expenditures in the United States were $183 billion, of which $177 billion could be attributed to expenditures within individual states, with the remainder falling under an undistributed, "other/unknown" category. The statistics and discussion below refer to state R&D levels in relation to the distributed total of $177 billion.
The "other/unknown" category includes R&D performed within the 50 states, or the District of Columbia, but where the specific location of such performance was not provided by survey respondents. It also includes R&D conducted by organizations within the United States, but where actual performance does not take place in a particular state or the District of Columbia, e.g., research conducted on marine vessels, and research in Puerto Rico.
For the compilation of these data, 15 sources were used, which include NSF statistical reports, as well as statistical reports from other Federal agencies, namely, the Department of Commerce (DOC), the Department of Labor (DOL), the Department of Education (DOE), and the U.S. Small Business Administration (SBA). A complete listing of these sources are provided at the end of this overview.
State Distribution of R&D Performance
R&D is substantially concentrated in a small number of states. In 1995, California had the highest level of R&D expenditures-over $36 billion-representing approximately one-fifth of the $177 billion U.S. total. The six states with the highest levels of R&D expenditures-California, Michigan, New York, Massachusetts, New Jersey, and Texas (in descending order)-accounted for approximately one-half of the entire national effort. The top ten states-adding, in descending order, Illinois, Pennsylvania, Maryland, and Ohio-accounted for nearly two-thirds of the national effort (figure1 and table 2). California's R&D effort exceeded, by nearly a factor of three, the next-highest state, Michigan, with $13 billion in R&D expenditures. After Michigan, R&D levels declined relatively smoothly to approximately $5 billion for Ohio. The 20 highest-ranking states in R&D expenditures accounted for about 85 percent of the U.S. total; the lowest 20 states, for 5 percent.
States that are highest in total R&D performance are usually ranked among the highest in industrial and academic R&D performance. For example, among the top 10 for total R&D, eight states were among the top 10 for industrial R&D, and eight were among the top 10 for academic R&D, as shown in table 2.
For Federal intramural research, there was less commonality with the top ten for total R&D. Only four states were found in both top-ten lists: Maryland, California, Ohio, and Texas. The six others in the Federal intramural list, in descending order of Federal R&D performance, were the District of Columbia, Virginia, Alabama, Florida, New Mexico, and Hawaii. Maryland ranked first among Federal R&D performers, followed by the District of Columbia, California, and Virginia. The placement of Maryland, the District of Columbia, and Virginia among the top four in Federal R&D performance reflects the concentration of Federal facilities and administrative offices within the national-capital area. Alabama, Florida, and New Mexico rank among the highest in Federal R&D because of their relatively high shares of Federal space- and defense-related R&D. Hawaii ranks among the highest in Federal R&D because of its relatively high level of intramural research conducted by the Department of Health and Human Services.
Ratio of R&D to Gross State Product
States vary widely in the size of their economies, owing to differences in population, land area, infrastructure, natural resources, and history. Consequently, variation in the R&D expenditure levels of states may simply reflect differences in economic size or the nature of their R&D efforts. A simple way of controlling for the size effect is to measure each state's R&D level as a proportion of its gross state product (GSP) (table 3). That proportion is referred to as R&D "intensity" or "concentration." Overall, the Nation's total R&D to gross domestic product ratio was 2.5 percent in 1995. The top 10 rankings for R&D intensity were, in descending order, New Mexico (8.1 percent), the District of Columbia, Michigan, Massachusetts, Maryland, Delaware, California, Connecticut, Rhode Island, and Idaho (the latter with an intensity of 3.5 percent). New Mexico's R&D intensity is largely attributable to Federal support to the Sandia National Laboratories and Los Alamos National Laboratory FFRDCs in the state, provided by the Department of Energy.
Figure 2 juxtaposes state R&D performance with GSP, with the 50 states and the District of Columbia ranked in descending order of R&D. R&D expenditures are displayed as a dark bar, measured on the upper axis; GSP is displayed as a wider gray bar measured on the lower axis; both are measured in billions of dollars. The two highest-ranked states in total R&D-California and Michigan-clearly show R&D levels that are relatively high in relation to their GSPs, as reflected by their presence in the top 10 list for R&D intensity (table 2).
Some states with below-average R&D intensity ranked high in total R&D performance because of their large economies. The state ranked third in R&D performance, New York, had a relatively low (1.8 percent) R&D intensity. Thus, its third-place position in total R&D performance may be a function of its large state economy. The same may be said of Texas, Illinois, Pennsylvania, Ohio, and Florida. In contrast, Massachusetts, New Jersey, and Maryland are more like California and Michigan, with relatively high R&D levels in relation to economic size. As can also be seen in figure 2, states with relatively low levels of total R&D tend, on average, to have low R&D intensity, with the exceptions of Delaware, Idaho, and Rhode Island. South Dakota, with the lowest total R&D level, also had the lowest R&D intensity (0.3 percent).
Federal Support for R&D
In addition to the performer-reported data described above, data on Federal support for R&D are available from surveys of the Federal agencies that provide such funds. Levels of Federal funding according to the surveyed performers can differ from levels according to the surveyed funding agencies. (See the technical note below on these differences.)
As reported by Federal agencies that fund R&D, the Department of Defense (DoD) and the Department of Health and Human Services (HHS) together provided 68 percent of the $67 billion in total Federal support for R&D to all types of performers in fiscal year (FY) 1995. California and Maryland were the two leading recipients of Federal R&D funds (table 4). Performers in California, primarily industrial firms, received 21 percent of DoD's R&D support. Maryland received 20 percent of HHS' funding, largely supporting intramural activities undertaken at biomedical research facilities at the National Institutes of Health (NIH). California received more R&D funds from both the National Aeronautics and Space Administration (NASA) and NSF than any other state. The main recipients in California of NASA R&D funding were industrial firms and FFRDCs, while the main recipients of NSF funding were universities and colleges. Maryland had the largest share of any one Federal agency's total R&D support, with one-third of the DOC R&D funds. Intramural research activities accounted for most of this funding, associated primarily with DOC's National Institute of Standards and Technology (NIST).
Differences in performer-reported and source-reported Federal R&D
The NSF collects, and these profiles contain, two separate estimates on total Federal funding of R&D. Survey data are obtained from both Federal funding agencies and performers of the work (Federal labs, industry, universities, and other nonprofit organizations). National totals, however, are based on data reported by performers because they are in the best position to (i) indicate how much they spent in the actual conduct of R&D in a given year and to (ii) identify the sources of their funds. Performer reporting also reduces the possibility of double-counting and conforms to international standards and guidance.
Historically, the two survey systems of sources and performers tracked fairly closely. For example, in 1980 performers reported using $29.9 billion in Federal R&D funding and Federal agencies reported total R&D obligations of $29.8 billion. In recent years, the two series have diverged considerably: For 1995, performers report $63.1 billion in Federal R&D support, compared with the $68.8 billion reported by Federal agencies (table 5). (Note that the $67.1 billion in Table 4 and in the U.S. total in the state profiles differs from the $68.8 billion amount because state data are collected from just 10 Federal agencies). The difference in the Federal R&D data totals appears to be concentrated in funding of industry. Overall, in each year since 1989, industrial firms have reported less in Federal R&D support than the amounts that Federal agencies have reported in supporting industrial R&D. The difference has been as large as $9.3 billion, observed in 1994. For 1995, Federal agencies reported $31.7 billion in total R&D obligations provided to industrial performers compared with $23.5 billion in Federal R&D funding reported by industrial performers (table 6). Consequently, data users are cautioned to exercise considerable care in comparing the R&D performance data in table 2 (and detailed in the upper half of the state profiles) with that reported by Federal agencies in table 4 (and detailed in the lower half of the profiles). NSF is investigating the causes of these divergent trends.
Science and Engineering Profiles
In addition to the state R&D statistics summarized above, the state profiles contain other statistics (from both NSF and non-NSF sources) relating to economic activity within the state. These data are included in a set of 52 one-page S&E profiles available in hard copy or from the World Wide Web. NSF survey indicators include numbers of doctoral scientists and engineers, doctorate degrees awarded by major S&E field, S&E graduate students and postdoctorates, amounts of Federal R&D obligations by agency and performer, total and industrial R&D expenditures, and academic R&D expenditures by major S&E field. Indicators from non-NSF sources include population, civilian labor force, per capita personal income, total Federal expenditures (not just on R&D), higher education expenditures, patents, small business innovation research (SBIR) awards, and GSP by originating economic sectors. In these profiles, state rankings and totals are provided for the 50 states, the District of Columbia, and Puerto Rico. Because data on total and industrial R&D expenditures are not available for Puerto Rico, rankings for those two variables exclude Puerto Rico.
Of the 17 main indicators ranked by state in the profiles (excluding the rankings in the bottom half of each profile involving Federal R&D obligations by state and performer), California ranked first in each except in personal income per capita, where it ranked 13th. New York ranked 2nd in eight of the indicators and ranked no lower than 8th in any of the others. Texas ranked between 2nd and 7th in all indicators, except in personal income per capita, where it ranked 32nd.
In this report, when states are ranked by a particular statistic, two or more states may have the same value for that statistic. When such ties occur, the tied states are given the same rank, and the next lowest state is given a rank equal to the number of higher ranked states plus 1. For example, if two states are tied for 27th place, they both receive a rank of "27," no state is given a rank of "28," and the next lowest state is given a rank of "29."
For many survey statistics used in this report, some fraction of the survey totals could not be allocated to specific geographic regions, or were for U.S. areas other than the 52 listed in this report (e.g., territories). Consequently, U.S. totals reported here may differ from those reported in the underlying surveys. Also, because of rounding, the sum of the gross state product sector percentages may not equal 100 percent.
For some states, reported levels of R&D expenditures and levels of doctoral scientists and engineers are relatively small. For these cases, sampling error in the surveys associated with these statistics may have bearing on the precision of these data, including state rankings. Particular caution in this regard should be used in comparisons among states with low levels of doctoral scientists and doctoral engineers. For example, South Dakota is ranked lowest in doctoral engineers with an estimated number of 77 in the state, and Wyoming is the next highest in rank with 117. However, according to the survey of doctorate recipients from which these data were obtained, any estimate of 100 doctoral engineers is subject to a standard error of 40, implying that the difference between these two states for this variable is not statistically significant. For 1,000 doctoral engineers, there is a standard error of 120. For doctoral scientists, the standard error for 100 scientists is also 40, and for 1,000 scientists it is 110. Readers should consult with the original sources of these data, as listed below, for additional information on standard errors associated with these and other statistics reported.
For information about, and copies of, Science and Engineering State Profiles, please contact:
Richard J. Bennof or Steven Payson
Research and Development Statistics Program
Division of Science Resources Studies
National Science Foundation
4201 Wilson Boulevard, Suite 965
Arlington, VA 22230
 In any discussion of R&D expenditures, an important distinction must be made between R&D "performance" (the situation in which R&D is actually carried out) and R&D funding "sources" (where the money for R&D originates). For example, a term like "Federal R&D" is ambiguous, in that it does not specify whether it is referring to performance or funding. The Federal Government is a much larger "source" of R&D funding (termed "Federal Funding of R&D") than a performer of R&D itself (termed "Federal Intramural R&D"). In the reporting of R&D by state, much more attention has been paid to R&D performance within states than R&D funding originating from states. Since R&D performance is an important component of the economic activity of the state, and the geographic location of funding organizations may have little bearing on economic activity within the same state, this report will focus on R&D performance.
 At present, data on R&D performed by nonprofit institutions within individual states include only R&D that derives from Federal funding. However, a survey of R&D by nonprofit organizations is now underway, which is expected to provide more complete data on R&D by nonprofit organizations in the near future.
 These complete data sets are only available in odd-numbered years. Thus, 1997 is the next year for such reporting and these data will be available by early 1999.
 Federal intramural performance includes the administration of extramural R&D programs.
 "Environmental Sciences" for S&E doctorate data are the sum of earth, atmospheric, and ocean sciences. "Life Sciences" for S&E doctorate data were defined as including both biological and agricultural sciences. Medical or health-related data are collected but non-S&E health fields are excluded.
 For three variables-personal income per capita, number of SBIR awards, and gross state product,-the data sources for Puerto Rico differ from those used to obtain state data.
 See "Methodological Report of the 1995 Survey of Doctorate Recipients," National Research Council, Washington, DC.