NISTIR 7280 - Identifying Technology Flows and Spillovers Through NAICS Coding of ATP Project Participants

DATA ANALYSIS

Large Increase in Number of Industries Covered by ATP Projects

As Jaffe (1996) notes, market and knowledge spillovers are more likely to occur when the technology may be used in many different applications, particularly technology that may be applied outside of developer's own-industry. The transition from three- to six-digit NAICS codes results in a large increase in the available detail of the ATP portfolio as measured by the number of different industries potentially affected by ATP-funded technology. Table 4-1 shows the increase in unique three- and six-digit NAICS codes before and after the coding. The number of unique six-digit own-industries is 106; the number of unique six-digit downstream industries is 180.

Distribution of Own-Industry for ATP Participants

Figure 4-1 shows the distribution of own-industries for ATP participants. The largest own-industry is NAICS code 541710, Research and Development in the Physical Sciences. This represents almost a quarter of all participants' own-industry categorizations.

Figure 4-1. Distribution of Own-Industry for ATP Participants by Six-Digit NAICS Code (Projects Started between January 1999 and July 2004)

As shown by figure 4-1, other own-industries include software, electrical equip­ment manufacturing and semiconductors. These industries may be characterized as midstream industries as shown on figure 3-1. They typically provide inputs to the upstream and downstream industries. This graph only provides the counts of the num­ber of firm's own-industry within each industry and is not adjusted for dollars spent per project or the fact that a single joint venture might contribute several companies to these counts while a single applicant provides only one. A disproportionate num­ber of ATP company participants are small firms, many of them are relatively new, and consider their primary business at this stage to be research and development. This fact explains the relatively large number of ATP firms whose own-industry is the R&D industry. As they develop commercial products, many of them will reclassify them­selves into the NAICS code of their new industry.

Distribution of Use-Industry for ATP Participants

An examination of figures 4-1 and 4-2 reveals differences between own- and use-industries. The own-industry category contains a significant number of R&D and software firms as well as instrument and machinery manufacturers (about 45 percent of total own-industry). Own-industries are likely to be characterized as midstream industries, which provide components, materials, and process technologies for both upstream and downstream industries.

Figure 4-2. Distribution of Use-Industry for ATP Participants by Six-Digit NAICS Code (Projects Started between January 1999 and July 2004)

Use-industries, on the other hand, include transportation equipment, medical care, computer manufacturing, and pharmaceutical manufacturing. Such industries are more likely to be characterized as downstream industries, which purchase components from midstream industries to assemble their products or deliver their services.

The "other" category in both figures 4-1 and 4-2 includes many NAICS codes too numerous to list here. However, Appendix B contains a list of all own and use-industry NAICS codes by commercial application.

Do ATP Participants' Own-Industries Generate Technology?

The Carnegie Mellon Heinz Business School conducted a study of NAICS codes and technology industries.¹ The researchers defined "technology employer industries" as those that employ three times the national average of employees in R&D, or almost 10 percent. They defined an industry as a "primary technology generator" if the industry per employee R&D expenditures exceeds the national average of $11,297 and if the industry's proportion of R&D scientists and engineers to total employees exceeds the U.S. average of 6.5 percent. A secondary technology generator industry meets one of these two criteria.

Using these definitions, we analyzed ATP participants' own- and use-industries to determine the percentage of firms that are in industries known as technology employers and either primary or secondary technology generators. Table 4-2 shows the results of this analysis.

An ATP firm's own-industry is more likely to be a technology employer and/or a primary or secondary technology generator than its use-industry. This finding suggests that ATP companies participate in technology-generating industries, but that the ben­efits go to users in industries less likely to be a significant generator of technology. It may be an indication that ATP catalyzes research by R&D performers whose spillover benefits may accrue to predominately non-R&D performers although further research would be necessary to answer this question.

Evidence of Multi-Use Technologies

Multi-use technologies possess high spillover potential. One way to measure such potential in ATP projects is to see whether project participants' own-industry differs from the use-industry of their proposed commercial applications.

Table 4-3 indicates that most ATP participants propose a commercial application outside of their own-industry. In fact, 84 percent of project participants propose at least one commercial application outside of their own-industry's six-digit NAICS code.

Potential diffusion of ATP-funded technologies is fairly consistent across the five technology areas that ATP uses to classify projects (listed in table 4-3); over 80 per­cent of all project participants within each technology area propose at least one application outside of their own-industry. Across all commercial applications, almost half are outside of the participant's own-industry.

Evidence of Infrastructural Technologies

Infrastructural technologies support the development of generic technologies and subsequent market applications, and may be described as "tools for the toolmak-ers." The first method used to determine infrastructural technologies was whether the proposed commercial application use-industry was the R&D industry (NAICS code 541710). Recall from figure 4-1 that almost a quarter of all own-industry participants were in the R&D industry. However, only 5 percent of use-industry applications are in the R&D industry.

We divided use-industry applications into the five ATP technology categories, and we found that biotechnology had the most use-industry commercial applications for the R&D industry. In fact, almost one out of five (17.8 percent) commercial applica­tions in biotechnology may be used by the R&D industry.

Advances in biotechnology tools drive biotechnology's R&D industry. In the past 10 years, more than 45 ATP awards have supported the development of diagnostic tools used to isolate and evaluate genetic information. Developments include produc­tion of a nucleic acid micro-array, a micro-fluidic system, an informatics package, and an integrated platform that offers faster and cheaper methods of producing genetic data on a routine basis.²

Do ATP Participants Intend to License ATP-Funded Technology Outside of Own-Industry?

Jaffe's third factor suggests the potential for spillovers is high when the proposed commercial application is outside of the own-industry and the commercial strategy is to license. Licensing provides an avenue for knowledge diffusion while improving the welfare of producers and customers. For example, a firm may develop a technology but lack the expertise or willingness to commercialize it outside of its own-industry. A license allows another firm to develop the commercial application for the industry; presumably, it costs less to license the technology than to develop it in-house. The firm that developed the technology receives a licensing fee, which defrays some of the costs of developing the technology, and provides an incentive for further development of enabling technology.³ Sometimes proposed commercial applications become impractical or unfeasible as the project progresses. Therefore, ATP queries companies each year about whether any proposed commercial applications are still viable. Table 4-4 presents an analysis of whether a commercial application was outside of the own-industry and whether those applications are still viable. An application is considered viable if the company indicated so in its last submitted BRS report (as of January 31, 2004).

The proportion of applications that are licensed outside of their own-industry is between 10 and 20 percent across the five technology areas. In four areas, at least 80 percent of the applications are still viable. The comparable proportion in manufactur­ing is somewhat lower: in this area, applications retain their commercial viability in a little over half of the time.

Do Applications Outside Own-Industry remain Commercially Viable?

Table 4-5 shows the extent to which proposed commercial applications remain viable as the project progresses. It shows that the proportion of applications reported as still viable since the last report is in the 90 percent range until project closeout; even then, the proportion only drops to slightly below 80 percent.

As Ruegg (1999) suggests, developing technology outside of one's own-indus­try can be difficult. We analyzed whether participants' commercial applications were outside of their own-industry and what percentage is still viable based upon each participant's latest submitted BRS report. Table 4-6 shows the data by technology area. The still-viable designation is based on the last report submitted involving each application.

Other Ways to Use the Data

Figures 4-3 and 4-4 illustrate the use of NAICS codes to explain how an ATP project diffuses technology from midstream to downstream industries. NAICS data are avail­able at the national, state, and county levels, so different levels of industry aggrega­tion can be selected to make this point. The example provided in figure 4-3 involves a titanium project based in Idaho. This project encompasses two applications, motor vehicle body and titanium powders, both of which industries are important to the Idaho economy. The project originates from a small research company specializing in titanium powders. If the project is successful, the benefits will flow into the Idaho economy through use by the motor vehicle body manufacturing sector, which generated $126 million in sales in 1997 for Idaho and the inorganic chemical manufactur­ing sector, which generated sales of $450 million and whose jobs pay an average of $52,000.

Figure 4-4 illustrates the potential benefits from a "Stirling engine" project involving Praxair Corporation. The company proposes three possible commercial applications: gasoline engines, air-conditioning and industrial gas manufacturing. The US markets for these industries are $18.6, $22.9 and $5.2 billion, respectively. Success in this project could potentially be diffused into markets totaling almost $50 billion employing over 180,000 people.

Figure 4-3. A Demonstration of Potential Economic Benefits to the State of Idaho Resulting from Spillovers of an ATP Project

Figure 4-4. A Demonstration of Potential Economic Benefits to the Nation Resulting from Spillovers of an ATP Project

Appendices and Discussion

Appendix A contains the list of Jaffe's factors affect market and knowledge spillovers. We analyzed three of these factors. We found evidence that ATP selects projects with the potential for broad economic impact. Multi-use technologies provide both market and knowledge spillovers. ATP enables multi-use technology development by either selecting single-applicant projects where the potential commercial applications are multiple and extend outside of the firm's own-industry or though selecting joint ventures such as the shrimp genetics' project involving multiple partners from a disparate set of own-industries; proposing commercial applications outside of their own-industry.

Other factors that lead to spillovers include projects where the commercial application involves research tools. We found that biotechnology is the technology area most likely to involve the use-industry research and development in the physical sciences; this is called an infrastructural technology. These may be also referred to as "tools for the toolmaker." As the Biotechnology Industry Organization states:⁴

As indicated by the above statement, biotechnology tool development has led to great spillover benefits outside of biotechnology and has spread throughout other sectors of the economy such as energy and agriculture. ATP contributes to this effort.

Appendix B contains the lists for each technology area of own- and use-indus­try counts by commercial application. These lists provide an overall view of the ATP portfolio and the potential impacts that ATP projects have on which industries. These lists may be used as an inventory of the areas that ATP enables at a given time and may be used with lists from other government programs in order to compare across the entire government's R&D portfolio, as was suggested by the 2002 Presidential Committee of Advisors and Science and Technology report, Assessing the U.S. R&D Investment.

____________________
1. Paytas and Berglund (2004).

2. www.atp.nist.gov/eao/2004annual/2004annual.pdf.

3. To the extent that the licensor possesses market power then the spillover benefit may be somewhat negated. But, a new technology may not be a perfect substitute for the existing technology, and therefore, the
licensor may have to reduce the price somewhat in order to induce the firm to purchase their license.

Return to Table of Contents or go to next section.

Date created: May 25, 2006
Last updated: June 7, 2006

ATP website comments: webmaster-atp@nist.gov  / Technical ATP inquiries: InfoCoord.ATP@nist.gov.

NIST is an agency of the U.S. Commerce Department
Privacy policy / Security Notice / Accessibility Statement / Disclaimer / Freedom of Information Act (FOIA) /
No Fear Act Policy / NIST Information Quallity Standards / ExpectMore.gov (performance of federal programs)