Technology Adoption Indicators Applied to the ATP Flow-Control Machining Project

NISTIR 6888
Technology Adoption Indicators Applied to the ATP Flow-Control Machining Project

3. Selecting a Case Study Using Technology Adoption Indicators (TAIs)

The primary mission of the National Institute of Standards and Technology (NIST) is to promote U.S. economic growth by working with industry to develop and apply technology, measurements, and standards. The Advanced Technology Program (ATP) at NIST supports this mission by providing cost share awards to industry to develop high-risk, "enabling" technologies which can ultimately increase economic growth, the quality of jobs, and the quality of life that comes from such growth.

ATP projects are designed to be partnerships between government and industry. Individual awards are made to single firms or joint ventures of firms to produce technologies that enable the development of new products, processes, and services across diverse application areas. Universities, state and federal laboratories, and other nonprofit institutions also participate in the projects as members of joint ventures and as subcontractors. Awards are made based on rigorous peer-reviewed competitions designed to select those proposals best qualified in terms of their cutting-edge technological ideas and potential for national economic benefits. Emphasis is placed on the difference that ATP funding will make. Awards have specific cost-share rules: for example, joint-venture participants pay more than half of the total project costs. Single-company awardees pay all indirect project costs and may also cover some of the direct costs. Each award has a set of specific goals and completion dates.

The Flow-Control Machining (FCM) Project is an ATP joint venture, its partners being Extrude Hone Corporation, Ford Motor Company, General Motors, the University of Pittsburgh, and the University of Nebraska at Lincoln. The four-year project began in 1996 and its total funding was $7.9 million; $4.0 million was provided by Extrude Hone, Ford, and GM, and the remainder was provided by ATP. The project goal was to develop two finishing processes that increase the functional precision of cast-metal parts that carry fluids in interior passageways. Its first targeted application was increased airflow in automobile engines, where the increase in functional precision can be used to increase engine horsepower, increase fuel efficiency, reduce emissions, and reduce the cost of engines. While these two processes were initially aimed at the automobile industry, they also have broad application in other, non-automotive sectors. If diffused into a wide array of industries, such as aerospace, manufacturing tooling, and medical, the impact could be quite large.

The particular capabilities of the FCM technology allow it to be applied in a number of industries with significant societal benefit. For example, using the processes in personal watercraft could significantly reduce noise and pollution.⁽¹⁶⁾ Using the FCM processes in lawnmower engines and other powered lawn and garden equipment could significantly decrease emissions and increase fuel efficiency and horsepower. Extrude Hone also indicated that airplane engine manufacturers could also be a potential adopter of the FCM technology. Using the processes in aircraft engine cooling systems could significantly reduce engine weight and therefore overall aircraft weight.

Of the many small-engine applications (including all-terrain vehicles and personal watercraft) that Extrude Hone was interested in, the company recommended the lawnmower engine application for this case study. A new EPA regulation, described in Section 3.4, was identified as a compelling driver of the adoption of the FCM technology in small engine applications. Interested manufacturers and Extrude Hone cooperated to test the FCM technology in lawnmower engines. Extrude Hone also indicated that airplane engine manufacturers could also be a potential adopter of the FCM technology, but was not pursuing that application.

Based on the interest shown by Extrude Hone and lawnmower engine manufacturers, as well as the economic implications of the use of the FCM technology in such a basic, widely used device, we selected the lawnmower engine application for an in-depth case study. We selected the airplane engine application as a comparison to the lawnmower application because the extremely high value and relatively low number of production units in airplane engine manufacturing provided a contrast to lawnmower engines.

In this chapter, the two applications-lawnmower engines and aircraft engines-illustrate the use of the TAI system as a screening mechanism to identify promising industries for technology adoption. This analysis results in the identification of lawnmower engines as the more likely adopter of the FCM technology. In the case study presented in the next chapter, we proceed with the quantitative analysis of the FCM technology in lawnmowers, based on our discussions with Extrude Hone and on the impact of a new EPA regulation, and supported by the TAI analysis.

In this chapter we illustrate the application of the TAIs, including the public policy (regulation) TAI.⁽¹⁷⁾ We use the TAIs developed in Section 2 to guide us in selecting the best application for a prospective case study from these choices. Section 3.1 addresses the concentration ratios and HHI index, section 3.2 joint ventures, section 3.3 the number of patents, section 3.4 public policy and regulation, and section 3.5 historical adoption trends. In the sections on industry concentration, joint ventures, and patents, we address the broadly defined measures (4-digit SIC and 6-digit NAICS industries) followed by the narrowly defined measures (specific product market data). The broadly defined comprehensive industry data are used to demonstrate the use of TAIs in comparing industries. Using narrowly defined specific industry data, we were able to more accurately calculate and apply the TAI measures for the selected industry. The impact of public policy and regulation on the industry selected for the case study is analyzed using industry-specific information.

Concentration ratio and HHI data were obtained from a table (accessed online) in the 1997 Economic Census, Manufacturing Subject Series (U.S. Bureau of the Census 1997): "Share of Value of Shipments Accounted for by the 4, 8, 20, and 50 Largest Companies in Each of the 3-, 4-, 5-, and 6-digit NAICS Industries: 1997." The four-firm concentration ratios (CR4) for all 6-digit (NAICS) manufacturing industries are shown in Appendix D. Figure 2 presents data on the CR4, Figure 3 on the CR8, and Figure 4 on the HHI in the two 6-digit industry groups that include lawnmower firms and aircraft engine firms.

For the CR4, we use the bounds of 40 and 60 to identify an industry with market concentrations optimal for technology adoption (following the analysis of the optimal bounds identified in Chapter 2). For the CR8, we use a range of 60 to 80 to define the optimal market concentration for technology adoption. The HHI may be interpreted in a similar fashion. We use the range of 1,000 to 1,800 as the optimal range for technology adoption.

Figure 2 . Four-Firm Concentration Ratios of Home Lawn and Garden Equipment and Aircraft Engines and Engine Parts

Figure 3 . Eight-Firm Concentration Ratios of Home Lawn and Garden Equipment and Aircraft Engines and Engine Parts

Figure 4 . Herfindahl-Hirschman Indices of Home Lawn and Garden Equipment and Aircraft Engines and Engine Parts

As shown in Figure 2 , the CR4 (64) for Home Lawn and Garden Equipment is above 60, the upper limit for optimal technology adoption. The industry is close to, though not quite within, the optimal CR4 range for technology adoption. The CR4 (77) result for Aircraft Engines and Engine Parts indicates the structure is not optimal because the top four companies hold too much market share to spur technology adoption. The CR8 for Home Lawn and Garden Equipment (80), shown in Figure 3 , is just barely within the optimal range of 60 to 80. However, the CR8 (82) for Aircraft Engines and Engine Parts is so close to the Home Lawn and Garden Equipment CR8 as to make any distinction impossible.

As shown in Figure 4 , the Home Lawn and Garden Equipment industry HHI is 1,172, which is within the optimal range. The HHI statistic, because it includes all the firms in an industry, is a more comprehensive and revealing measure of industry concentration. Even though the CR8 for the two industries is similar, the HHI reveals that the Home Lawn and Garden Equipment industry is less concentrated than the Aircraft Engine and Engine Parts industry. We conclude that this industry consists of a large number of small firms that are large enough to have the resources and motivations to adopt new technology. In contrast, the industry that includes manufacturers of aircraft engines greatly exceeds the upper bound with an HHI statistic of 2,058. This signifies that the industry is more highly concentrated than is optimal for the adoption of new technology.

We applied the same measures (CR4, CR8, HHI) to narrowly defined CR4, CR8, and HHI data developed for the lawnmower industry itself.⁽¹⁸⁾ Specialized data on the lawnmower market were obtained from the Power Systems Research OELink database (Power Systems Research 2001). These data allowed the creation of CR4, CR8, and HHI exclusively for lawnmower manufacturers. Figures 5 and 6 show the CR4 and CR8 concentration ratios developed from the OELink data. Figure 7 shows the HHI developed from the OELink data.

Constructed from the narrowly defined industry data, the CR4 and CR8 concentration indices both exceed the optimal ranges for technology adoption, throughout the entire time range. The more comprehensive and revealing HHI index, however, remains well within the optimal range for technology adoption. The HHI, while exhibiting a sharp spike mid-decade, ended at 1,666 in the year 2000, which is well within the optimal 1,000 to 1,800 range.

The concentration indices presented in Figures 5 to 7 are likely to be overestimates of the true indices for two reasons. First, the OELink data define the relevant marketplace narrowly. The inclusion of companies that could but are currently not producing lawnmowers lowers all the concentration indices. Other companies that use or produce small engines include all-terrain vehicles, personal watercraft, outboard motors, small motorcycles and mopeds, and snowmobiles. Overestimation of concentration indices also results because both imports of finished lawnmowers with U.S. engines and U.S.-assembled lawnmowers with non-U.S.-manufactured engines are not counted.

We conclude that, on the whole, the concentration ratio measures provide stronger support for adoption by the lawnmower industry than the aircraft engine industry. The comprehensive HHI measure indicates a strong preference for the lawnmower industry.

This section examines patent data to identify whether the two industries show a proclivity to supply technology. We describe the patent data source and the application of the patent TAI to both lawnmower and airplane engine manufacturers. Then the patent indicator is analyzed within the Structure-Conduct-Performance (SCP) framework, and the likelihoods of technology adoption by the two different manufacturers are compared.

The analysis of the number of patents in an industry is used to indicate the "supply" of innovation available to that industry. In this study, we are not only concerned about the adoption of technology in general, but with the adoption of a specific technology. All else being constant, the fewer the number of competing technologies, the greater the likelihood of any one technology innovation being adopted. Specifically, if there is a great demand for technological innovation but few sources of technological innovation, then a new innovation has a greater chance of adoption than if there were numerous innovations.

The mechanism by which the number of patents influences technology adoption is less theoretically founded than the connections of technology adoption with concentration ratios and HHI indices. Consequently, the best method for measuring the precise number and timing of patents is unclear, as is the interpretation of the selected measure. In this report, we use the total number of patents granted over the last five years.⁽¹⁹⁾ Patent count data were obtained from a U.S. Patent and Trademark Office publication, Patenting Trends in the United States 2000 (USTPO 2001). The USPTO has created a concordance between the U.S. Patent Classification System (USPCS) and a limited number of SIC codes. For the two industries under evaluation, patent data are available at the 3-digit SIC level.

Lawnmower engines are part of SIC code 3524, which is available at the 3-digit resolution (352) in the Patenting Trend database. Aircraft coolant systems are part of SIC code 3724, which is available at the 3-digit resolution (372) in the Patenting Trends database. Data from the last five years were summed to arrive at the total number of patents in the last five years (1996 to 2000, inclusive). This was done to indicate the current "supply" of relatively recent innovations in the most closely affiliated industry.

Figure 8 shows the whole and fractional patent counts for Farm and Garden Machinery (SIC 352), and Aircraft and Parts (SIC 372). The whole and fractional counting methods were described in Chapter 2 (section 2.2.1). The Farm and Garden Machinery industry has only two-thirds as many patents as the Aircraft and Parts industry, under the whole patent counting method. Under the fractional counting method, both industries have a similar number of patents.

Appendix B displays the whole and fractional patent counts in the manufacturing sector industries with SIC codes 30 to 39. Both lawnmower and aircraft engine manufacturers are in industries with a moderate level of patent activity. A substantial number of industries have worse performance.

This indicator reveals a preference for the lawnmower case study candidate using the whole patent counting method, and supports the selection of the lawnmower industry. Under the fractional patent counting method, it does not show a clear difference between the two candidates, and therefore does not lend support to either.

Figure 8 . Farm/Garden Machinery and Aircraft Patents Issued 1996 to 2000

The USPTO provides guidelines for manually constructing patent counts at a higher resolution. These guidelines involve searching for text that matches the selected product or industry in the patenting system Internet database. After collecting all relevant patent classification numbers, the number of patents in each of these categories can be collected.

A multi-year survey period reflects the supply of innovations available for "supply-push" adoption in a particular industry. Narrowly defined data on patents related to lawnmowers were obtained from the USPTO database and then sorted into two categories: patents relating to lawnmower engines and patents relating to lawnmowers in general.⁽²⁰⁾ Both types of patents are included because some lawnmower manufacturers make their own engines. The results over the past 6-year period are shown in Figure 9 .

Figure 9 shows that over the last six years, lawnmower-related patents increased from 24 to 35 per year, then declined to 28. The number of engine-related patents increased from 3 to 6 per year. From 1996 to 1998, 11 patents were issued. The next three years, 1999 to 2001, saw 19 patents issued. Reasons for this increase could include new emissions regulations, which are discussed in following sections. The figures for 1999 to 2001 indicate a recent increase in the technology "supply" available to lawnmower engine manufacturers. That increase supports the observation that suppliers of new technology target applications toward industries with high technology demand and low supply. In this case, Extrude Hone (which developed the FCM technology for automotive applications) could apply the FCM technology in an industry (such as the lawnmower industry) with a demand for improved engine technology. This is an example of a new "crossover" application for a previously developed technology.

The analysis of the number of RJVs in an industry is used to indicate the "supply" of innovation available to that industry. In this study we are concerned with the possible adoption of a specific technology (the FCM technology). All else being constant, the fewer the number of RJVs, the fewer the number of resulting technologies. It follows that fewer competing technologies enhance the likelihood of any one technology innovation being adopted. Specifically, given demand for technology, if there are few innovations, then a new innovation has a greater chance of adoption than if there were numerous innovations.

This section describes the source of the RJV data and the application of the RJV as an indicator of technology innovation supply to both lawnmower and airplane engine manufacturers. Then the likelihoods of technology adoption by the two different industries are compared.

RJV data were obtained from the Collaborative Research (CORE) database.⁽²¹⁾ This database includes the SIC code of each RJV filing in the Federal Register, by date.⁽²²⁾ Up to two SIC codes (at the 2-digit level) are available for each joint venture. We constructed a whole count and a fractional count, following the procedure outlined in the section on patents. A whole count adds up each occurrence of an SIC code-i.e., a RJV reporting two SIC codes would be counted once in each classification. A fractional count adds only one-half to each SIC code when the RJV reports two classifications. Appendix C shows RJV formation for all manufacturing industries at the 2-digit SIC level.

Lawnmower engines are included under SIC 35, "Machinery, Except Electrical." Aircraft engines are included under SIC 37, "Transportation Equipment." As shown in Figure 10 , the RJV indicator reveals that SIC 35 (50 and 42) has substantially fewer RJVs than SIC 37 (131 and 116), using the whole and fractional RJV counting methods, respectively.

Given highly aggregated data at the 2-digit level, this information may not be directly applicable to narrow industries, such as lawnmower engines. Nevertheless, this TAI reveals that the industry that includes lawnmowers is comparatively underserved by RJVs. The RJV indicator therefore supports the selection of the lawnmower industry. More narrowly defined TAI data on RJV formation by lawnmower manufacturers could not be found, but future case studies may be able to utilize narrowly defined RJV data.

Comparing government regulation across industries presents conceptual complexities. Public policy, such as government regulation, tends to have unique impacts and effects on industries. The impact of a regulation depends on, among other things, the type of regulation and its severity. Some regulations mandate changes in a currently marketed product, and they differ in the degree of change required in the product; other regulations effectively prohibit certain product innovations. Perhaps due to these conceptual hurdles, there is currently no broad attempt to measure regulation across industries. Regulation effects need to be examined on an industry-by-industry basis.

In this study, we found that the lawnmower manufacturers were constrained by pending Environmental Protection Agency (EPA) regulations intended to reduce small-engine emissions by 59% by 2007. The EPA Phase 1 and Phase 2 regulations are described in detail in Appendix A. The EPA expects that its regulations will cause all remaining side-valve (SV) engines used in lawnmowers to be converted to overhead-valve (OHV) engines. Some existing OHV engines used in lawnmowers will also require improvements to meet the standards. No pending regulations affecting technology adoption by aircraft engine manufacturers were identified.

In the above case, the EPA regulations will, by their design, strongly encourage small-engine manufacturers to make engine innovations and to adopt new emissions-reducing technologies, but it may not be clear in some cases what level of effort is required by the firms to meet the regulations. This illustrates the difficulties encountered when trying to make cross-industry comparisons of the effects of regulations on adoption. Many regulations, such as the above regulations on lawnmower emissions, contain text listing explicit goals mandated (e.g., number of lawnmowers with lowered emissions), providing the needed data for more detailed intra-industry analysis of the regulation's effect on technology adoption.

The SCP model takes into account historical, industry-specific trends for all TAI measures, and includes the history of past technology adoption. There is no standardized way to account for past technology adoption across industries. Each analysis requires both specific interpretations of what constitutes technology adoption and time-trend data on those observable characteristics.

In the case of lawnmowers, EPA Phase 1 emissions regulations in effect as of 1997 led to the replacement of some SV engine configurations with cleaner OHV engine configurations. The market penetration of the more sophisticated engine technology-such as OHV engine configurations-may be used to proxy technology adoption.⁽²³⁾ The Power Systems Research OELink database (Power Systems Research 2001) provides data from 1992 to 1999 on engine kW rating, stroke, valve cam, valves per cylinder, cycle, cylinders, displacement, configuration, and torque, all of which are subject to technological improvement. We use these data to analyze technology adoption trends for the lawnmower engine industry.

Figure 11, using yearly production data from Power Systems Research, shows the decrease in SV engines and increase in OHV engines, and indicates that lawnmower manufacturers adopted OHV engine technology coincident with Phase 1 emissions regulations. It is likely that Phase 2 EPA regulations will have a similar effect, increasing the adoption of new engine technology.

On the whole, TAIs based on the structure-conduct-performance model show that lawnmower manufacturers are more likely than aircraft manufacturers to adopt the FCM technology spillover from the automotive industry. The broadly defined TAIs are summarized in Table 1, and the narrowly defined, industry-specific TAIs are summarized in Table 2.

Table 2. Summary of Findings Using Narrowly Defined Data on Lawnmower Manufacturers

Measure	Lawnmower Engines
CR4 and CR8	Too concentrated to adopt
HHI	Optimal size for adoption
RJVs	No specific data
Patents	Recent slight increase indicates some adoption, probably due to regulation
Public policy/regulations	High likelihood of adoption due to major new regulation
History of technology adoption	Indicates some adoption, probably due to regulation

Overall, the TAI analysis suggests that the likelihood of adoption by the lawnmower industry is strong. The broadly defined market concentration indicators show a clear preference for the lawnmower manufacturing application over aircraft engine application. The narrowly defined lawnmower industry data show that, according to the four- and eight-firm concentration ratios, the lawnmower industry is too highly concentrated to adopt technology. However, the Herfindahl-Hirschman Index, which is a more comprehensive measure than the CR4 and CR8, supports technology adoption by the lawnmower industry. The broadly defined patent TAI supports the selection of either the lawnmower industry or the airplane engine industry. The broadly defined research joint venture TAI shows that the lawnmower industry was underserved by RJV formation, compared to the airplane industry, suggesting an opening for an external, new technology. The narrowly defined TAI data on patents, public policy, and historical technology adoption show that Environmental Protection Agency regulations are likely to encourage the lawnmower manufacturers to adopt new engine technology.⁽²⁴⁾

Figure 11 . Side-Valve and Overhead-Valve Market Penetration, 1990 to 2000

_____________________

16. Recent tests indicate that FCM can decrease personal watercraft (PWC) emissions by 50% while increasing power
by 16%. Extrude Hone Corporation, “News Bulletin,” October 13, 2000.

17. The industry-specific measures were developed after all the broadly defined measures had been considered and a
decision made on which application to choose for a case study. Once the target industry—lawnmowers—had been
selected, additional narrowly defined data were developed. For presentation purposes, we appended the specific,
narrowly defined measures immediately after the discussion of the broadly defined measures.

18. Once the target industry—lawnmowers—had been selected (based on all the broadly defined TAIs), additional
narrowly defined data were developed. In order to present these additional measures in the most appropriate
location, we appended the specific, narrowly defined measures immediately after the discussion of the broadly
defined measures. The specific measures were developed after all the broadly defined measures had been considered
and a decision made on which application to choose for a case study.

19. The Census data do not indicate the length of patent maintenance during the 5 year summary.

21. Collaborative Research (CORE) database, 2001. The CORE database is funded by the National Science
Foundation (NSF) and maintained by Dr. Albert Link of the University of North Carolina.

22. The current CORE database contains 8,185 records, of which 763 report SIC codes. Unfortunately, identification
of the target SIC code is not required by the Federal Register. The coding of each filing was done manually after the
fact.

24. Regulations can drive an industry to adopt technology not previously targeted to that industry. The need to meet
the emissions regulation required an evaluation of all engine technology innovations, including those originally
targeted at automobile engines.