Chapter 8 | Invention, Knowledge Transfer, and Innovation

Innovation Indicators: United States and Other Major Economies

Intangibles in the economy include many services, such as insurance, education, telecommunications, as well as experiences such as concerts, movies, and sporting events; brand images; and embedded technology, such as software in cars and nutritionally enhanced food products (Blair and Wallman 2001).

Some intangibles once created provide benefits for years to come, for example computer software, R&D activity, designs and artistic creations. Often they can be simultaneously in more than one location, adding a dimension of use that tangibles do not possess. Digitization also allows many types of intangibles to be transmitted digitally across networks, multiplying potential impact further.

Gross domestic product (GDP) statistics for many countries, including the United States, include investment measures for the following types of intangible capital: computer software and databases, R&D expenditures, and artistic originals. Artistic originals are long-lived artwork produced by artists, studios, and publishers, including music, books, and programming, as measured by the Bureau of Economic Analysis (BEA) (Soloveichik and Wasshausen 2013) and tabulated by the U.S. Bureau of Labor Statistics as part of its measurement of productivity. Figure 8-14 and Figure 8-15 show investment for the U.S. manufacturing sector and for the nonfarm, nonmanufacturing sector in computer software, R&D, and artistic originals, adjusted for inflation with 2009 as the base year. The data are based on the categories used in the national income accounts. To prevent double counting in these measures, R&D directed toward the creation of computer software is categorized with computer software rather than with R&D (BEA 2013).

Outside of manufacturing, the relative magnitudes of R&D and computer software investment differ and software investment comprises a much larger share of overall investment in intangibles. In the nonmanufacturing sector, investment in computer software in 2015 ($282 billion) is more than six times as large as that of the manufacturing sector ($41 billion) (Figure 8-15). Additionally, investment in artistic originals is considerably larger than in R&D. In 2015, investment in these artistic originals was $78 billion. Digitization and networking allow these originals to be transformed into downloadable and streaming services; such services are increasingly consumed using personal devices such as laptops, tablets, and cell phones.

The indicators presented cover important, but not all, types of intangible capital. Some firm investments are in the human capital embedded in people. Formal investments in education, training, and health; and experience gained through on-the-job training and other activities may be not only capital for the individual but also for the firm. A broader perspective on intangible capital suggests that all investments in intangibles that firms use repeatedly over time should be treated as capital assets. Other types of activities that could be included as intangible capital include spending on designs, spending to develop and protect brands, spending to develop human capital in the firm, and spending devoted to organizational development (Corrado, Hulten, and Sichel 2005, 2007).

Access to financing is an essential component of the translation of inventions to innovations. Entrepreneurs seeking to start a new firm to commercialize a nascent or emerging technology rely on several funding sources: the entrepreneur’s own funds, friends and family, bank loans, venture capital, angel investment, and government support (OECD 2014:174). Patterns and trends in venture capital investment are an indicator of support for emerging technologies that could make their way into the economy or are increasing their use in the economy. Venture capital investment is also an important financing source for existing high-technology firms that are commercializing technology. This section uses data from PitchBook, a company that collects comprehensive global data on venture capital and other early-stage investment.

Venture capital investment is generally categorized into three broad stages of financing—seed stage, early stage, and later stage. Seed-stage financing supports proof-of-concept development and initial product development and marketing and is important for understanding emerging technology trends. Global seed-stage venture capital investment was $6 billion in 2016, accounting for a very small share (4%) of total venture capital investment (Figure 8-16; Appendix Table 8-30). Early-stage financing accounted for 40% ($52 billion) of total venture capital investment in 2016. Early-stage financing supports product development and marketing and the initiation of commercial manufacturing and sales (Figure 8-16; Appendix Table 8-30); it also supports company expansion and provides financing to prepare for an initial public offering (IPO). Later-stage financing accounted for 56% ($73 billion in 2016) of total venture capital financing. Later-stage financing includes acquisition financing and management and leveraged buyouts. Acquisition financing provides resources for the purchase of another company, and management and leveraged buyouts provide funds to enable operating management to acquire a product line or business from a public or a private company.

Venture capital has been highly concentrated in early- and later-stage financing over the last decade and a half (Figure 8-16). The limited amount of seed-stage financing has been attributed to the reluctance of venture capitalists to invest in the uncertain and risky state of new product development (World Bank 2010:90). The difficulty of entrepreneurs obtaining seed-stage financing contributes to the “valley of death,” the inability of new and nascent firms to obtain financing to commercialize their inventions and technology (OECD 2014:174).

Global seed-stage venture capital investment was $5.8 billion in 2016 (Figure 8-17; Appendix Table 8-30). The United States received $3.3 billion, the largest share (58%) by far of any region or country. The EU and Israel were the second and third largest recipients, receiving $0.9 billion and $0.7 billion, respectively.

Global seed-stage investment has grown exponentially over the last decade, from more than $300 million in 2006 to $5.8 billion in 2016 (Figure 8-17 and Figure 8-18; Appendix Table 8-30); its growth rate (34% annualized average rate) was more than double that of early- and later-stage investment (13% annualized average rate), resulting in the seed-stage share of total investment increasing from 1% to 4% (Figure 8-18). Despite its strong growth over the last decade, seed stage remains a very small share of total venture capital investment.

In the United States, seed-stage financing grew from less than $200 million in 2006 to $3.3 billion in 2016 (Figure 8-17; Appendix Table 8-30). Like global trends, it grew far more rapidly (34% annualized average) than total U.S. early- and later-stage investment (9% annualized average).

PitchBook classifies firms that receive venture capital investment by industry (Appendix Table 8-31). Venture capital-backed firms that operate in multiple industries are classified in multiple industries. Classifying firms in multiple industries gives a more comprehensive picture compared with single industry classification because many firms produce products and services in multiple and diverse industries. The disadvantage is that the sum of venture capital investment in multiple industries exceeds total investment because of the double counting of investment in companies that are classified in multiple industries.

Between 2011 and 2016, the two industries that received the largest amount of seed-stage investment were software as a service ($3.8 billion) and mobile ($3.5 billion) (Figure 8-19; Appendix Table 8-32). Industries that received between $0.8 billion and $1.1 billion were financial technology, e-commerce, big data, and artificial intelligence. (Big data consist of companies that provide a product or service that is too large for traditional database systems.) Artificial intelligence, consisting of a variety of technologies, including software, natural language processing, and optical character recognition technology that has close ties to science, received $0.8 billion. Life sciences, a technology that is also closely tied to basic research in biotechnology and pharmaceuticals, received $0.5 billion.

Industries that have rapid increases in seed-stage investment may indicate nascent or emerging technologies. Investment in autonomous cars went from zero in 2013 to $74 million in 2016 (Figure 8-20; Appendix Table 8-32). Investment in the early- and later stages also grew very rapidly in this industry (Figure 8-21). (See the discussion in section U.S. early- and later-stage venture capital investment by industry.) Robotics and drones had the largest increase from 2013 to 2016 in investment (139% annualized average rate), reaching $175 million in 2016. Investment in virtual reality grew at a 91% annualized average rate to reach $72 million; early- and later-stage investment also rapidly increased (Figure 8-21). Artificial intelligence grew by a 74% annualized average rate between 2013 and 2016 to reach $347 million, the highest level of investment among fast-growing industries in 2016. Investment in the Internet of Things was robust (56% annualized average rate), reaching $141 million. (See Chapter 6 sidebar The Internet of Things for a discussion of these technologies.) Investment in life sciences, an area that includes pharmaceuticals and biotechnology that is closely linked to basic science, grew more slowly (39% annualized average rate between 2013 and 2016), also reaching $141 million.

Global early- and later-stage venture capital investment was $125 billion in 2016 (Figure 8-22; Appendix Table 8-30). The United States attracted the most investment ($65 billion) of any region or country, accounting for slightly more than half of global investment. China attracted the second largest amount of investment ($34 billion) with a global share of 27%. The EU attracted the third largest amount ($11 billion).

Between 2006 and 2013, early- and later-stage global venture capital investment remained annually in the range of $30–$60 billion before surging to $99 billion in 2014 (Figure 8-22; Appendix Table 8-30). After increasing by 31% to $130 billion in 2015, investment fell slightly to $125 billion in 2016 because of high valuations of venture-backed companies, the lack of exits of existing venture-backed firms, and political and economic uncertainties (KPMG 2017:7).

Investment in China soared from $3 billion in 2013 to $34 billion in 2016, the largest increase of any country (Figure 8-22). China’s share of global investment climbed from 5% in 2013 to 27% in 2016. The rise of China’s middle class with disposable income and the government’s focus on promoting domestic innovation have prompted major investments by private venture firms based in China and other countries, largely from the United States. The Chinese government has also created almost 800 public-backed venture capital funds that have raised more than $300 billion to invest in China (Oster and Chen 2016).

Early- and later-stage venture capital investment in the United States rose sharply (63%) to reach $65 billion between 2013 and 2016. Despite the robust growth in U.S. investment, the U.S. global share dropped from 69% in 2013 to 52% in 2016 due to more rapid growth in China. One factor that has driven the growth in U.S. investment is Chinese-based venture capital investors who have invested heavily in U.S. startups and venture-backed firms; one source estimates that about one-quarter of venture capital invested in the United States in 2015 originated from China (Oster and Chen 2016). China’s growing wealth and the government’s push to develop innovative high technologies have prompted Chinese-based companies and wealthy individuals to invest in U.S. startups and acquire technology (Dwoskin 2016).

In other regions and economies, investment in the EU rose from $6 billion in 2013 to $11.0 billion in 2016 (Figure 8-22; Appendix Table 8-30). After spiking from $1.4 billion in 2013 to $7.7 billion in 2015, investment in India fell to $3.3 billion in 2016, more than double its level in 2013 (Figure 8-23). The spike in venture capital funding has been due to several factors, including the election of the first single-party government in 30 years, strong macroeconomic fundamentals, India’s focus on S&T in higher education, and the country’s strong position and expertise in business services and e-commerce.^​

Investment in Israel more than doubled from $0.7 billion in 2013 to $1.7 billion in 2016 (Figure 8-23; Appendix Table 8-30). The expansion of venture capital outside of the United States, particularly in China, coincides with the globalization of finance, greater commercial opportunities in rapidly growing developing countries, and the decline of yields on existing venture capital investments in U.S.-based companies.^​

Between 2011 and 2016, software as a service ($81 billion) and mobile ($68 billion) received the largest total amount of early- and later-stage venture capital investment (Figure 8-24; Appendix Table 8-33). These two industries also received the largest amounts of seed-stage investment during this period. The next two largest were life sciences ($43 billion) and e-commerce ($35 billion)—the former being a technology that is closely tied to basic science. Four industries—lifestyles of health and sustainability, manufacturing, financial technology, and clean technology—each received $20 billion to $23 billion. Big data received comparatively less early- and later-stage investment ($18 billion), although it ranked high in seed-stage investment (Figure 8-19).

Rapidly growing early- and later-stage investment in industries may be an indication that these areas are maturing and moving from radical or transformative to more incremental technological change. Between 2013 and 2016, ephemeral content, technologies that provide online sharing and temporary display of photographs and other content, had the most rapid growth in investment (193% annualized average) among all industries, soaring from $74 million to $1.9 billion (Figure 8-21; Appendix Table 8-33). More than 20 companies, including Snapchat, Instagram, and Periscope, have received venture capital financing for this rapidly growing sector.^​ Venture capital and other investors sold Snapchat to the public in a IPO in March 2017 (Balakrishnan 2017).^​

Investment in virtual reality grew the second fastest (104% annualized average rate), rising from $164 million to $1.4 billion (Figure 8-21; Appendix Table 8-33). Autonomous cars had the third fastest increase (102% annualized average), jumping from $56 million to $459 million. More than 70 companies, including Tesla, Mobileye, and Delphi Automotive, have received venture capital financing in this sector to develop software, computers, cameras, radar sensors, and other technologies.^​ Most major automakers are conducting pilot tests of autonomous cars or have made large investments in or acquisitions of companies with autonomous driving technologies (Gates et al. 2016).

Investment in three-dimensional printing increased from $86 million to $612 million (92% annualized average). Lifestyles of health and sustainability grew the sixth fastest (79% annualized average rate), from $1.5 billion to $8.7 billion in 2016. (Lifestyles of health and sustainability consists of companies that provide consumer products or services focused on health, the environment, green technology, social justice, personal development, and sustainable living.) Early- and later-stage investment in artificial intelligence and machine learning, which has rapidly growing seed-stage investment, rose from $1.2 billion in 2013 to $3.9 billion in 2016.

Between 2011 and 2016, mobile technology was the leading industry receiving early- and later-stage investment ($37 billion) in China (Figure 8-25; Appendix Table 8-34). This industry received the second largest investment in the United States (Figure 8-24). E-commerce was the second largest ($19) in China and the fourth largest in the United States. Software as a service was the third largest, receiving $15 billion; this industry received the most investment in the United States. Life sciences, a technology that is closely tied to basic science, was the fifth largest, receiving comparatively little investment ($2 billion) compared with the four leading industries. This industry was the third largest in the United States, receiving far more investment ($43 billion) than in China.

Starting in the late 1970s, concerns by national policymakers about the comparative strength of U.S. industries and their ability to succeed in the increasingly competitive global economy took on greater intensity. The issues raised included whether the new knowledge and technologies flowing from federally funded R&D were being effectively exploited for the benefit of the national economy, whether pervasive barriers existed in the private marketplace that worked to slow businesses in exploiting new technologies for commercial applications and implementing innovations, and whether better public-private partnerships for R&D and business innovation had the potential to enhance the nation’s economy to respond to these emerging challenges (Tassey 2007). There was also a concern about how to avoid inappropriately placing the government in positions to substitute for private business decisions better left to the competitive marketplace.

Many national policies and related programs have been directed at these challenges over the last 30 years. One major national policy thrust has been to enhance formal mechanisms for transferring knowledge arising from federally funded and performed R&D (Crow and Bozeman 1998; National Research Council [NRC] 2003), a topic discussed in the chapter’s previous section. Another important development has been clearer recognition by policymakers, entrepreneurs, and the investment capital sector that structural and market barriers—often termed technological and commercial “valleys of death”—can arise in the marketplace that create difficult-to-bridge gaps for the innovation process and all too many barrier-filled pathways for otherwise promising new technologies (Branscomb and Auerswald 2002; Jenkins and Mansur 2011). These insights and an associated set of diagnostic concepts have given rise to several government programs intended to address the main sources for the gaps, with the intent of strengthening the prospects for the development and flow of early-stage technologies into the commercial marketplace. Other policy initiatives have included a particular focus on accelerating the commercial exploitation of academic R&D and encouraging the conduct of R&D on ideas and technologies with commercial potential by entrepreneurial small and/or minority-owned businesses.

The sections immediately following focus on this second theme of the commercial exploitation of federally funded R&D and review status indicators for several significant federal policies and programs directed at these objectives.

The Small Business Innovation Research (SBIR) program and Small Business Technology Transfer (STTR) program are longstanding federal programs that provide competitively awarded funding to small businesses for purposes including stimulating technological innovation, addressing federal R&D needs, increasing private-sector commercialization of innovations flowing from federal R&D, and fostering technology transfer through cooperative R&D between small businesses and research institutions. The U.S. Small Business Administration (SBA) provides overall coordination for both programs, with implementation by the federal agencies that participate (SBA 2015).

The SBIR program was established by the Small Business Innovation Development Act of 1982 (P.L. 97–219) to stimulate technological innovation by increasing the participation of small companies in federal R&D projects, increasing private-sector commercialization of innovation derived from federal R&D, and fostering participation by minority and disadvantaged people in technological innovation. The program has received several extensions from Congress since then and is presently authorized through 2017. Eleven federal agencies participate in the SBIR program: USDA, DOC, DOD, the Department of Education, DOE, HHS, DHS, the Department of Transportation, the Environmental Protection Agency, NASA, and NSF.

The STTR program was established by the Small Business Technology Transfer Act of 1992 (P.L. 102–564, Title II) to facilitate cooperative R&D by small businesses, universities, and nonprofit research organizations and to encourage the transfer of technology developed through such research by entrepreneurial small businesses. Congress has likewise provided several extensions since it was initially enacted, with the program continuing through 2017. Five federal agencies participate in the STTR program: DOD, DOE, HHS, NASA, and NSF.

For SBIR, federal agencies with extramural R&D budgets exceeding $100 million annually must currently (FY 2017) set aside at least 3.2% for awards to U.S.-based small businesses (defined as those with fewer than 500 employees, including any affiliates). (The set-aside minimum was 2.5% for FYs 1997–2011, rising incrementally to 2.9% in FY 2015, 3.0% in FY 2016, and 3.2% in FY 2017.) Three phases of activities are recognized. In Phase I, a small company can apply for a Phase I funding award (normally not exceeding $150,000) for up to 6 months to assess the scientific and technical feasibility of an idea with commercial potential. Based on the scientific and technical achievements in Phase I and the continued expectation of commercial potential, the company can apply for Phase II funding (normally not exceeding $1 million) for 2 years of further development. Where the Phase I and II results warrant, the company pursues a course toward Phase III commercialization. The SBIR program itself does not provide funding for Phase III; depending on the agency, however, Phase III may involve non-SBIR-funded R&D or production contracts for products, processes, or services intended for federal government use. Several agencies offer bridge funding to Phase III and other commercialization support for startups (NRC 2008:208–16).

The initial round of SBIR awards was for FY 1983. This yielded 789 Phase I awards, across the participating agencies, for a total of $38.1 million of funding (Table 8-7; Appendix Table 8-35 and Appendix Table 8-36). The scale of the program expanded considerably thereafter. To date, the number of awards peaked in FY 2003, when the annual total of awards was 6,844 (5,100 Phase I awards and 1,744 Phase II awards). The peak in funding to date was FY 2010, with total funding of $2.300 billion ($565 million for Phase I awards and $1.735 billion for Phase II awards). More recently, however, the annual number of awards and funding totals have dropped somewhat (Table 8-7). In FY 2015, the award total was 4,508 (2,939 Phase I awards and 1,569 Phase II awards), with total funding of $1.923 billion ($462 million for Phase I awards and $1.461 billion for Phase II awards). In FY 2015, most funding reflected awards by DOD (49%) and HHS (22%) (Appendix Table 8-36). DOE (10%), NASA (8%), and NSF (7%) accounted for smaller shares. The other six participating agencies were 1% or less of the total.

For the STTR program, federal agencies with extramural R&D budgets that exceed $1 billion annually must currently (FY 2017) reserve not less than 0.45% for STTR awards to small businesses. (The set-aside minimum was 0.3% for FYs 2004–11, rising incrementally to 0.4% in FYs 2014–15 and to 0.45% in FY 2016 and thereafter.) STTR operates within the same three-phase framework as SBIR. Phase I provides awards for company efforts to establish the technical merit, feasibility, and commercial potential of proposed projects; the funding in this phase normally does not exceed $100,000 over 1 year. Phase II is for continued R&D efforts, but award depends on success in Phase I and continued expectation of commercial potential. Phase II funding normally does not exceed $750,000 over 2 years. Phase III is for the small business to pursue commercialization objectives, based on the Phase I and II results. The STTR program does not provide funding for Phase III activities. Furthermore, to pursue Phase III, companies must secure non-STTR R&D funding and/or production contracts for products, processes, or services for use by the federal government.

The STTR program started with a single Phase I award for $100,000 in FY 1995 (Table 8-7; Appendix Table 8-35 and Appendix Table 8-37). This program has also expanded considerably in subsequent years. The peak years to date for number of awards were FY 2004, with a total of 903 awards (719 Phase I awards and 184 Phase II awards), and FY 2010, also with 903 awards (625 Phase I awards and 278 Phase II awards). The peak in total funding was $299 million in FY 2010 ($79 million for Phase I and $220 million for Phase II). In FY 2015, 725 awards were made (553 for Phase I and 172 for Phase II), with funding totaling $258 million ($99 million for Phase I and $159 million for Phase II). Fewer federal agencies participate in STTR, but those dominant in SBIR are also dominant in STTR. STTR awards from DOD accounted for 49% of the $258 million award total in FY 2015 (Appendix Table 8-37). HHS accounted for 24% of the STTR awards, and the remaining awards were from DOE (10%), NASA (9%), and NSF (8%).

The federal policies, authorities, and incentives established by the Stevenson-Wydler Technology and Innovation Act (and the subsequent amending legislation) and the SBIR and STTR programs are far from the whole of federal efforts to promote the transfer and commercialization of federal R&D. Many programs for these purposes exist in the federal agencies. These programs typically have objectives that closely reflect the specifics of agency missions and draw resources at levels well below the federal-wide SBIR and STTR programs. Several of the larger programs currently run by federal R&D performing agencies are briefly described in Table 8-8. A larger group of such federal agency policies and programs is documented in Appendix Table 8-38. Following Table 8-8, commentary is offered on three particularly well-known programs: DOC’s Hollings Manufacturing Extension Partnership, DOE’s Advanced Research Projects Agency-Energy, and NSF’s Industry/University Cooperative Research Centers.

The Hollings Manufacturing Extension Partnership (MEP) is a nationwide network of manufacturing extension centers located in all 50 states and Puerto Rico. MEP was created by the Omnibus Foreign Trade and Competitiveness Act of 1988 (P.L. 100–418) and is headed by DOC’s NIST (DOC/NIST 2017). The MEP centers (which are nonprofit) exist as a partnership among the federal government, state and local governments, and the private sector. MEP provides technical expertise and other services to small and medium-sized U.S. manufacturers to improve their ability to develop new customers, expand into new markets, and create new products. The centers work directly with manufacturers to engage specific issues, including innovation and business strategies, product development and prototyping, lean and process improvements, workforce development, supply chain development, technology scouting, and transfer. The centers also serve to connect manufacturers with universities and research laboratories, trade associations, and other relevant public and private resources. The MEP annual report for FY 2015 (the most recent report presently available) describes the national network of MEP centers as operating with a total budget of about $300 million annually—$130 million from the federal government (with more than $110 million going directly to the centers), with the balance from state and local governments and the private sector (DOC/NIST 2015). The MEP report indicates that technical expertise and other services were provided during FY 2015 to 29,101 U.S. manufacturing companies and attributes impacts of $8 billion in increased or retained sales, 68,477 jobs created or retained, and $1.2 billion in cost savings for these businesses. (These services and impact metrics are comparable with the reports of recent previous years.)

DOE’s Advanced Research Projects Agency–Energy (ARPA-E) provides funding, technical assistance, and market development to advance high-potential, high-impact energy technologies that are too early-stage for private-sector investment (DOE 2017). The main interest is energy technology projects with the potential to radically improve U.S. economic security, national security, and environmental quality—in particular, short-term research that can have transformational impacts, not basic or incremental research. The America COMPETES Act of 2007 (P.L. 110–69) authorized ARPA-E, and it received $400 million of initial funding through the American Recovery and Reinvestment Act of 2009 (P.L. 111–5). Federal funding (appropriations) for ARPA-E was $180 million in FY 2011, $275 million in FY 2012, $251 million in FY 2013 (reduced by budget sequestration that year), and $280 million in FYs 2014 and 2015. ARPA-E’s annual report for FY 2015 (most recent available) indicated 81 new project awards in FY 2015, with a total of 542 funded projects and $1.49 billion of funding since the program’s inception (DOE 2015). The program identifies 31 focused and 2 open project areas, with topics including advanced batteries, transportation technologies, solar energy, energy storage technologies, advanced carbon capture technologies, electric power transmission, distribution and control, biofuels, and improved building energy efficiencies.

NSF’s Industry/University Cooperative Research Centers (IUCRC) program supports industry-university partnerships to conduct industrially relevant fundamental research, collaborative education, and the transfer of university-developed ideas, research results, and technology to industry (NSF 2017). NSF supports IUCRC through partnership mechanisms where, per NSF, the federal funding is typically multiplied 10–15 times by supplementary funding from businesses and other nonfederal sources. The IUCRC program report for 2015–16 (NSF/IUCRC 2017) indicates 68 centers across the United States, with more than 1,000 nonacademic members: 85% are industrial firms, with the remainder consisting of state governments, national laboratories, and other federal agencies. NSF’s IUCRC program funding for the centers was about $17.2 million that year, with other sources of support (including NSF funds other than the IUCRC program; member fees; funds from industry; and funds from other federal agencies, state government, and other nonfederal government), bringing the total of center funding that year to $109.3 million. Research is prioritized and executed in cooperation with each center’s membership organizations.

The data presented thus far on invention, knowledge transfer, and innovation provide insights into the sources of knowledge, inventions, and funding for innovation, as well as the efforts by government and academic institutions to facilitate technology transfer and the early-stage development of useful technologies. Yet none of these measures provide a clear indicator for the incidence of innovation in firms—the implementation of a new or significantly improved product or business process. Firm-level survey data collected in the United States, Europe, and parts of Latin America, Asia, the Pacific, and South Africa provide industry-level data on the incidence of innovations, as well as rich ancillary data on related activities in firms. See sidebar Concepts and Definitions for Business Innovation Survey Data for more information on the framework behind the U.S. survey data on innovation collected by the National Center for Science and Engineering Statistics BRDIS. This U.S. survey is also the source of the R&D expenditure data reported in Chapter 4.

Per NSF’s BRDIS, 17% of U.S. firms (or companies) reported introducing a new or significantly improved product or process during 2013–15 (Table 8-9): 1 in 6 firms. This incidence rate of innovation varies across firm size and industry. Reported innovation rates increase overall with firm size. However, across all firms, more than 230,000 that have fewer than 250 employees (and at least 5) had introduced a product or process innovation. For large firms, those with 250 or more employees, more than 6,500 introduced innovations.

ICT-producing industries report many of the highest rates of innovation in manufacturing and in other sectors of the economy. Within manufacturing, almost half of electronic equipment and component firms, and more than half of computer and electronic products firms reported innovations between 2013 and 2015 (Figure 8-26). Outside of manufacturing firms, 44% of computer systems design firms and 31% of information industry firms reported innovations (Figure 8-27).

Overall, one-third of manufacturing firms reported an innovation, accounting for more than 37,000 firms with innovations. Firms in paper (34%), plastics and rubber (38%), and petroleum and coal products (38%) report innovation rates above one third. For chemicals, transportation equipment, and miscellaneous manufacturing, the industry innovation incidence rates are higher yet—more than 40%.

Outside of manufacturing, 15% of firms, or 200,000 firms, reported innovations. In addition to the ICT-producing industries discussed earlier, transportation and warehousing, health care services, electronic shopping and auctions, and scientific R&D services, among others, have incidence rates above the nonmanufacturing average (Figure 8-27).

Focusing on product innovation compared with process innovation, manufacturing firms overall report product and process innovations at similar rates, about one-quarter of firms. For nonmanufacturing firms, these rates are about 1 in 10. Across industries, U.S. firms reported higher rates of process innovation compared to product innovation (Table 8-10).

Interest in international competitiveness drives cross-country comparisons of business innovation rates, and these indicators provide a uniquely focused measure of activity distinct from R&D.

The data described as follows are collected under The Oslo Manual (OECD/Eurostat 2005), discussed in the sidebar Concepts and Definitions for Business Innovation Survey Data. While differences in survey methodologies across countries continue to drive inconsistency among international data, broad patterns emerge. Across countries, the highest rates of product and process innovation are reported in relatively smaller, but S&T-focused economies, such as Switzerland, Israel, and Finland. In contrast, Japan, the United Kingdom (UK), and the United States all rank relatively low in reported incidence (Table 8-11).

Not surprisingly, country-level data show innovation incidence varies across firm size. Firms with 250 or more employees had higher innovation rates than smaller firms, with a notable exception. For Australia, small firms had a higher product innovation rate compared with larger firms.

Cross-national comparability complicate interpretation of the OECD innovation data. The subjective element in respondent identification of something “new or significantly improved” can vary systematically across countries, and may miss incremental improvements. Also, U.S. survey data identify only new or significantly improved products and processes, whereas Community Innovation Survey data include separate categories for organizational innovation and marketing innovation. Industry and firm size coverage also varies across countries for the surveys.

Statistical agencies have primarily focused their attention on business-sector activity. However, inventors and entrepreneurs have long played an important role in innovation. Individual innovators invent, implement, and share innovations, whether as a tool or as a hobby. Both kinds of activities generally fall outside the scope of business innovation surveys.^​

Less well understood than business innovation, improvements in collaborative tools and Internet connectivity increase the importance of individual innovators (Gault and von Hippel 2009). Academic researchers in the United States, the UK, Japan, Finland, and South Korea gathered information on free innovation by households between 2012 and 2015, focusing on new product development and modifications (von Hippel 2017). Although relatively small scale (fewer than 2,000 respondents for the United States and the UK each), these surveys find household innovation rates between 1.5% for South Korea and 6.1% for the UK.^​

Although this activity is less well understood than business innovation, improvements in collaborative tools and Internet connectivity increase this activity’s importance (Gault and von Hippel 2009). A design, computer program, or set of instructions, for example, can be shared for free through the Internet, allowing free reuse throughout the world. Teams of connected contributors add to potential impact.

Innovations contribute to economic growth through cost savings from new and improved processes and from sales from new products. New knowledge about the innovation also spreads through the economy. New firms enter and competitive forces can shift the composition of output to higher-productivity firms. If this impact is sufficiently large, we might expect to see rising growth in the ratio of quantity of goods and services produced by workers (GDP) relative to hours worked, measured as labor productivity. Figure 8-28 shows U.S. labor productivity averages for four subperiods between 1990 and 2016. Overall, productivity growth in the United States has been on a declining trend since the early 2000s, including during the economic recovery after the Great Recession (Figure 8-28).

Many factors in addition to the impact of innovation contribute to productivity, including workforce skill and investments in physical and intangible capital. As an indicator of the impact of innovation on economic growth, productivity can be decomposed into component parts, where multifactor productivity is the part attributed to technology’s overall impact on the economy. It is calculated as the output growth that cannot be attributed to labor and capital inputs, after accounting for changes in workforce skill and the quality of capital. Figure 8-28 shows that trends in MFP in the United States have been similar to trends in labor productivity: MFP grew faster on average between 1995 and 2007 compared with the first half of the 1990s, and growth moderated since 2007.

The moderation in the growth rate of MFP is evident in other developed economies as well, including France, Great Britain, and South Korea (OECD 2017). Figure 8-29 shows MFP and GDP growth for the 10 largest OECD countries for two periods: 2001 to 2007 and 2009 to 2015. For each country, the height of each bar is GDP growth. In addition to MFP, increases in labor and increases in capital used in the economy contribute to growth. The factors are shown in Figure 8-29 within each bar: labor input, capital input, and MFP. Only Germany had more than a nominal increase in overall productivity growth across these periods. For Germany, increases in labor and MFP contributed to the growth. For Japan, MFP contributions to growth offset smaller contributions from capital.

More broadly, MFP growth has been depressed in both developed and developing economies since the global financial crisis of 2008. Lingering effects of the global recession may be responsible. Structural factors remaining from the recession include corporate debt ratios, misallocation of capital within and across sectors, slower ICT investment, and shifting preference toward less risky investments (Adler et al. 2017).

Many explanations for the slowdown focus on the pace of innovation and technology diffusion from ICT investment. Gordon (2016) argues that the period of the late 1990s to mid-2000s was one of unusually rapid growth from the spread of Internet-enabled communications, entertainment, and commerce, and that the future pace of innovation is unlikely to match this period. From this perspective, MFP is in a secular slowdown, with the gains from investment in ICT in the late 20th century having ended, and the major innovations of the late 19th and early 20th centuries were not and are unlikely to be followed by innovations that have as significant an effect on MFP growth.

An alternative explanation is that MFP growth may be delayed by lags between innovation and its systemic diffusion and adoption (Brynjolfsson and McAfee 2014). Historically, such delays have been especially prominent for general purpose technologies (GPTs; see sidebar General Purpose Technologies), a special category of technologies that are widely used, capable of ongoing technical improvement, and enable innovation in application sectors (Bresnahan 2010).^​

Measurement issues also effect the clarity of MFP as an indicator of the impact of innovation, since MFP is measured as a residual from economic data. High-quality expenditure data on inputs and outputs are necessary, and supplementary measures needed for good measurement are quantity, price, depreciation, and rate-of-return data for capital (Hall and Jaffe 2012).^​

The policy implications for the apparent productivity slowdown are large, motivating better understanding of the causes of the slowdown at the level of individual firms. Changes in firms can be obscured by aggregate sector statistics. The data best suited to explore these dynamics are firm-level data (e.g., those available in the U.S. Census Bureau’s Business Dynamics Statistics). These data provide information on establishments opening and closing, firm startups and shutdowns, and their associated employment impacts. The data show that business dynamism, as measured by new startup formation, has been declining in the last decade, leading to fewer firms and older firms (Decker et al. 2014). Importantly, since 2000, the number of high-growth young firms has declined.

Based on U.S. Census data, half of U.S. firms were 5 years old or younger in 1982; this share has steadily declined, reaching 32% in 2014 (Figure 8-30). Along with this decline in the share of young firms, there have been corresponding steady decreases in the share of new job creation and in the share of overall employment from young firms. Young firms accounted for 19% of employment in 1982, and the share declined to 10% by 2014. Although most startups fail and most of the startups that do survive do not grow, a small share of these fast-growing firms makes a disproportionately large contribution to job growth (Decker et al., 2014).

Although the factors behind these trends are not well understood, industry concentration and barriers to entry for inventors and entrepreneurs may be factors contributing to this decrease in dynamism in the U.S. economy. Foster and coauthors (2017) suggest that career paths of entrepreneurs and the activity of new firms are areas in which better data and analysis can help explain how innovation activity affects productivity.