Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Chapter V Dimensions and Distribution of R&D Total Magnitude of Effort The character and distribution of research and development activity in the United States is difficult to describe or to explain in a brief survey. Although large volumes of data have been assembled on certain aspects, many features of the national distribution of R&D are not clearly perceived because of the great complexity of the objectives, sponsorship, and locational considerations involved. While a historical perspective sometimes helps to identify the forces that led to the creation of such complexes as those centered in Boston or Los Angeles, there are also many features of such developments that were fortu- itous. Furthermore, the presentation of the distribution of R&D according to geographic boundaries often fails to portray certain key aspects of the nature of the R&D enterprise or the features of the supporting infrastructure of a community that are conducive to the development of such enterprise. A brief review of some of the available data will indicate some of the patterns of the present composition and distribution of R&D. This should help to provide insights into the role that federal policies may have played in furthering either the concentration or dispersion of R&D activity. The analysis that follows may also indicate some of the constraints that are encountered in any attempt to alter federal R&D policies. In 1965, total expenditures for R&D in the United States were $20.5 billion, of which about $14 billion was carried out in industry, approximately $3 billion in federal laboratories, and the remainder divided between univer- sities and other non-profit institutions. Of these funds, the major source was 50
SI the federal government, which contributed $13 billion. Industry received al- most $8 billion; thus, the federal government supported somewhat more than half of all industrial R&D. In terms of manpower, the magnitude of the national R&D effort in 1965 consisted of 504,000 full-time-equivalent R&D scientists and engineers. Of these, industry employed about 351,000, or 70 percent of the total; 69,000 were employed by the federal government; 55,000 in universities and colleges proper; and 28,500 in federal contract research centers and other non-profit institutions. Some over-all perspective is useful; the total number of R&D scientists represented less than one percent of the total labor force of 75 million in that year, while the total expenditures for R&D represented about three percent of the gross national product of $681 billion. Perhaps as significant as the total magnitude of effort in R&D have been the rates of growth of effort or expenditures in recent years. These have varied between sectors and have shown a recent tendency to level off. However, in the period between 1945 and 1965, the rate of growth in manpower was much greater than in most other areas of national effort. For example, R&D manpower increased (over the period as a whole) at a rate in excess of seven percent a year, while the total civilian labor force grew at less than two percent a year. However, it is probably more meaningful to compare this rate of growth with that of the "professional and technical" category of the work force, of which the scientists and engineers constituted 16.5 percent of the total (in 1963). The scientist and engineer work force increased 91 percent from 1950 to 1963, compared with an 85 percent increase for the professional and tech- nical category as a whole (140 percent for scientists, 75 percent for engineers). Thus the growth in scientific and technical manpower was part of a larger trend in which the proportion of professional personnel as a whole increased very rapidly in comparison with the total work force. Between 1945 and 1965 R&D expenditures rose by almost 13 percent a year, while the gross national product increased by less than four percent a year (in real terms). During much of this period, federal funding of R&D grew at a substantially greater pace than did the industrial commitment. However, during the past five years, the privately funded industrial R&D has grown at a greater rate, while the rate of growth of federal funding has tended to level off, under pressures of congressional and administrative limitations on ex- penditures. Total federal obligations for R&D in the fiscal year 1968 were approximately $17 billion or about two percent of the gross national product. Distribution by Function and Performer The National Science Foundation categorizes R&D activities as basic research, applied research, or development. Basic research, which accounts for about
52 one eighth of total R&D expenditures, comprises original investigations for the advancement of scientific knowledge that do not have specific commer- cial objectives. Applied research work, which receives almost one fourth of R&D funds, is defined as investigations that are directed to the discovery of new scientific knowledge and that do have specific commercial objectives with respect to products or processes. Developmental activities, which re- ceive over three fifths of all R&D outlays, are technical activities concerned with translating research findings or other scientific knowledge into products or processes. Research work is widely distributed between the public and private sectors and within each of them. Colleges and universities typically perform the largest share of basic research work in the United States, but less than half. Governmental and industrial laboratories also perform significant amounts of basic research, as do private research institutions. Industrial firms and govern- ment laboratories both carry on substantial applied-research programs. A smaller scale of activity is reported by educational institutions and other non- profit organizations. In contrast, industrial firms do the great bulk of all development work, about three fourths in recent years. The remainder is accounted for mainly by government agencies and private non-profit research institutions. While gov- ernment turns to industry for the bulk of its development work, it does fund sufficient development work in universities to lend a more development- oriented flavor to university research than it would have in the absence of government funding. It is noticeable that the use that universities make of their own funds for support for R&D is much more highly oriented toward basic research than are the funds that they receive from government. Undoubt- edly many in the universities conceive of their principal R&D missions as being contributions to basic knowledge, and view any forces pushing them in the direction of a development orientation as weakening or undermining their fundamental missions. Private support for R&D comes from a wide variety of sources, including business firms, philanthropic foundations, and individuals. The public support, in contrast, has been highly concentrated in recent years. Two federal agen- cies, the Department of Defense and the National Aeronautics and Space Ad- ministration, account for four fifths of the federal financing. Most of the remainder is spent by two other agencies, the Atomic Energy Commission and the Department of Health, Education, and Welfare. Distribution of R&D by Industry As indicated in Table 1, a sizable fraction of the federally sponsored R&D is carried on in three major industry groups that receive a major portion of their
S3 R&D funds from the federal government: (1) aircraft and missiles, (2) electri- cal equipment and communications, and (3) scientific instruments. These in- dustries, whose combined sales represent only 17 percent of the total of U. S. industry, perform 85 percent of the federally sponsored R&D. However, these three industries account for only 20 percent of the company-sponsored R&D. In contrast, the rest of American industry, which received the remaining 15 percent of federally sponsored research and development, financed the bulk of company-sponsored R&D, spending on the average almost three dol- lars for each dollar received from federal sources. Some of the differences are undoubtedly associated with the maturity of the industry; other differences have to do with patterns of procurement in defense, space, and other areas. Although the research-intensive industries play a far more than average role in technological advance and innovation, they afford job opportunities to only a small fraction of the industrial population. The three previously identified research-intensive industries employ 56 percent of the 350,000 sci- entists and engineers engaged in industrial R&D; however, they provide job opportunities to only 2.8 million, or about 16 percent, of the 17 million em- ployees in manufacturing industries. That is, high-technology industry pro- Table 1 Distribution of R&D by Industries (1965) Scientists and Engineers per 1,000 Employees (thousands) Federal R&D Funds ($ millions) Company R&D Funds ($ millions) Total Employment (thousands) Aircraft and Missiles 4,500 620 113 873 Electrical Equipment and Communications 1,978 1,189 S3 1,676 Professional and Scien- tific Instruments 125 261 36 295 Chemicals and Allied Products 190 1,187 40 1,022 Transportation Equip- ment 326 913 20 1,244 Machinery 258 870 27 1,228 Primary Metals 8 209 5 1,074 Fabricated Metal Products 17 129 15 442 Other Manufacturing Industries 102 957 10 3,374 Non-Manufacturing Industries 255 103 10 936 Total, R&D-Performing Industries 7,759 6,438 30 12,164 Source: National Science Foundation.
54 vides job opportunities preferentially for the most technically skilled portion of the labor market. Relative to the aggregate number of job opportunities, this sector of industry makes a relatively small direct contribution to the total. The interactions between these industries as related to technology and in- novation can not be assessed from such tables, however. Many of the manu- facturing processes in firms that are not R&D-intensive are greatly affected by technology (both hardware and software) developed in industries that per- form major segments of federally sponsored and company-sponsored research. For example, the airline industry neither finances nor conducts large amounts of research and development. Yet the services it provides embody a large in- put of science and technology. The airlines, of course, rely on the aircraft and manufacturing companies for the R&D work on aircraft, and on manufacturers of computers, radars, and communications equipment for R&D relating to the management of reservations and air-traffic control. Similar relationships exist in the case of scientific instruments and electronic equipment, which are uti- lized by many manufacturing companies that themselves do comparatively little R&D. Thus, although technologically sophisticated products and ser- vices may represent a very small percentage of the value added in a given in- dustry, they may be indispensable to productivity. Hence the percentage of the gross national product represented by sales of high-technology products may be a very misleading measure of true economic importance. Moreover, the products of high technology may indirectly create many new job opportunities in other industries. The concentration of R&D activities in a relatively small number of organi- zations is noteworthy. This factor also will bear on the geographic analysis that follows. The greater part of research and development activity is carried out in institutions that employ 1,000 or more R&D scientists or engineers. About 80 percent of industrial R&D is carried out by the 100 largest firms, while over 90 percent of university research is carried out by the 100 largest universities. (50 percent is carried out by 24 universities.) It is to be noted, however, that company concentration exceeds university concentration. The top four companies in R&D performance accounted for 21 percent of all R&D and 28 percent of all federal R&D in 1965. The leading nine companies accounted for about half of federal R&D, whereas a compara- ble fraction of federally supported academic R&D involves 25 universities. The relation between R&D and firm size is susceptible to misinterpretation. When firms are ranked according to the size of their R&D expenditures, they give the appearance of a very high concentration of R&D performance in large firms. However, as shown by Comanor59 and by Scherer,60 if firms are also ranked according to net sales, it is found that R&D performance is proportional to net sales. The percentage of net sales invested in R&D may vary widely from
55 industry to industry but this proportionality holds for almost all types of in- dustry. This is true whether specified in terms of expenditures, number of patents, or number of technical employees. For example, the largest eight firms in sales accounted for 15.6 percent of sales, 14.6 percent of R&D employ- ment, and 9.3 percent of patents. On the other hand, the first eight companies ranked in order of R&D performance accounted for 35 percent of R&D and only 11 percent of net sales. The statistical paradox arises because different in- dustries spend widely differing percentages of net sales on R&D, and the science-intensive industry classifications tend to spend more on R&D. How- ever, within a given classification, there appears to be no bias toward higher percentages of R&D for large firms. Geographic Distribution of R&D In recent years, probably more emphasis has been placed on the question of the geographic distribution of governmental R&D funds than on any other as- pect of federal policies regarding R&D. Table 2 shows the regional distribution of R&D and related variables in 1965. It is immediately apparent that the distribution indicates certain significant disparities. It has often been brought into public view that the states of Cali- fornia, New York, and Massachusetts, with a combined population of 22 per- cent of the national total, perform 46 percent of federally supported R&D. By contrast, the states of the Midwest (the East North Central, West North Central, and West South Central regions), with 38 percent of the population, carry out only 17 percent of the federally supported R&D. A feature of geographic distribution of R&D not often emphasized is that company-sponsored R&D has a pattern of geographic distribution distinctly different from that which is federally sponsored. Thus the northeastern part of the country (including the New England, Middle Atlantic, and East North Central regions), which accounts for 29 percent of federal obligations, per- formed 72 percent of company-sponsored industrial R&D. In contrast, western regions (Pacific, Mountain, and West South Central) received 55 percent of the federal R&D contracts going to industry and accounted for only 16 percent of company-sponsored R&D. The geographic distribution also shows significant differences between the categories of research and the performing sectors. As shown in Table 2, R&D funds in educational institutions are more uniformly distributed geographi- cally than are federal obligations to industry (with the exception of New England). Much of the concentration of graduate research in Boston and New York is historical; this area provided a preponderance of graduate de- grees prior to World War II, and the support of academic research is signifi-
e o 5 I o. Ill â W O so ex CA O ca *Â«? 3 I VO O\ Â§ â¢a a; o 3 A a o I ob a cS <s V "S a H 3 c 2 5 S.8 IP Â« U oo 3 c o â¢ft ^H (N -H -H -H o e M Â£ ii , i Ill lis i â¢o e Â§ T> e 2Â§ o o ii *- 1) 83 Â«a II 56
57 cantly related to the number of advanced degrees conferred. A very substantial diffusion in terms of geographic distribution has occurred since 1946. The large universities of the Midwest perform a substantial portion of the graduate research and grant a sizable and increasing fraction of the advanced degrees in the nation. Another dimension of the geographic allocation of federal research and development funds is the extent to which the funds are received by the high- income or low-income areas. Such comparisons should not be pushed too far because, as discussed elsewhere in this report, the long-term impacts of R&D are not necessarily felt in the regions in which the work is done. Neverthe- less, such comparisons can provide useful insights. Table 3 arrays the eight regions of the United States, ranging from the Far West, with the highest per capita income, to the Southeast, with the lowest average income level. It can be seen that, in general, the high-income regions tend to receive shares of R&D contracts larger than their proportion of either population or income and that, conversely, the low-income regions tend to receive below-average shares. However, the patterns are not uniform. In the case of NASA programs, for example, the southeastern region received a larger share of these R&D funds than of the total personal income, and the Mideast a substantially smaller share. An important feature of the geographic distribution of scientists and engi- Table 3 Regional Distribution of Population, Income, and Federal Financing of Research and Development in Fiscal Year 1964 Percentage Distribution Personal Income Federal Research and Development Obligations Region Population Defense NASA NSF High Income 34.0 39.4 65.5 61.2 44.9 Far West 12.7 14.8 38.0 47.6 14.8 Mideast 21.3 24.6 27.5 13.6 30.1 Average Income 36.1 37.5 19.3 13.4 38.2 New England 5.8 6.5 7.2 1.9 13.0 Great Lakes 19.7 21.1 6.2 3.4 15.8 Plains 8.2 7.7 1.7 7.8 4.3 Rocky Mountains 2.4 2.2 4.2 0.3 5.1 Low Income 29.9 23.1 15.2 25.4 16.9 Southwest 8.1 6.8 7.9 6.6 10.1 Southeast 21.8 16.3 7.3 18.8 6.8 Grand Total 100.0 100.0 100.0 100.0 100.0 Source'. National Science Foundation.
58 neers is that the R&D scientist works in an urban setting. It has been estimated that 25 percent of all scientists are employed in seven metropolitan areasâ New York City, Washington, Los Angeles, San Francisco, Boston, Chicago, and Philadelphia. The manpower data available do not distinguish between scientists and engineers who work on R&D and the total grouping of scien- tists and engineers. A sizable majority of the engineering population (about 65 percent) is committed to professional tasks other than research or develop- ment. However, a survey61 of 14 major metropolitan districts in the United States shows that these cities contain 26 percent of the total population, 37 percent of the scientists, and 41 percent of the engineers. But it should be noted that, while these cities contain 27 percent of the total population, they contain 38 percent of the urban population. In other words, scientists and engineers do not usually live on farms or in small towns; for example, while 50 percent of the general population of Massachusetts lives in the Boston area, 63 percent of the state's scientists and engineers live there. The Chicago area has 62 percent of Illinois' population, and 74 percent of the scientists and engineers. It follows that, in sparsely populated states, there may be very few centers for research and development, partly because there are few centers of popula- tion. In the mountain states, which generally view themselves as "have nots" in this field, there are relatively few metropolitan areas; however, Denver and Salt Lake City carry out important shares of the nation's R&D. If research and development tend generally to be located in metropolitan areas, this is particularly true for small R&D firms, especially those that have recently come into existence. While large corporate centers for research may be relatively self-contained (for example, General Electric, Boeing, and DuPont), the newly formed high-technology corporation derives great benefit from sharing services and business derived from location close to other high-technol- ogy industry. These considerations should be taken into account in any ap- praisal of the geographic distribution of R&D in the nation at large.