|
|
||||||||||||||||||
|
|
|
NISTIR 7280 - Identifying Technology Flows and Spillovers Through NAICS Coding of ATP Project Participants CREATION OF THE NAICS DATA SETData SourcesThe ATP Economic Assessment Office (EAO) uses multiple survey instruments, collectively referred to as the BRS, to capture project participants' business data and commercialization progress; these data help in evaluating the success of ATP projects.1 Between 1993 and 1998, EAO used a disk-based survey instrument that asked companies to provide the SIC codes applied to their potential commercial applications.2 Beginning in 1999, EAO switched to a web-based survey instrument and to the collection of NAICS codes. In 1999 and 2000, companies were asked to identify the three-digit NAICS codes for both their own-industry and the industry of their potential business applications. Starting in 2001, EAO reevaluated this request, when it determined that it placed a large reporting burden on the companies but did not yield particularly helpful or consistent data.3 Box 3-1 defines the SIC and NAICS system of classification. Box 3-2 outlines important attributes of NAICS. Using this focal information as guidance, EAO assigned six-digit NAICS codes to project participants' own- and use-industries (see box 1-1). Potential spillovers and other associated economic impacts from an ATP project can be more precisely measured using six-digit NAICS codes than the three-digit level. For example, NAICS code 325, which represents all chemical manufacturing, contains 34 separate six-digit industries. By assigning six-digit NAICS codes to own and use-industry, these data will better enable ATP to trace the technology flows that potentially result from ATP projects.
These data help reveal how ATP-funded technologies are diffused across multiple industries. One method used to track spillovers created by a technology is to identify the own-industry and the use-industry. The own-industry is the primary industry in which the company operates; the use-industry is that of the potential business application(s) resulting from the new technology. Detailed industry data facilitate research into the pathway by which the benefits developed by a company from its own-industry spill over into the downstream industries of the various business applications. If ATP were to rely on three-digit NAICS codes alone, the spillover potential would be seriously understated. Simple Model of Economy FlowsFigure 3-1 shows a simple model of the economic pathways by which different industries produce goods for each other (inter-industry flows) and then how raw materials and intermediate goods end up in their ultimate final use by the consumer. The figure uses three types of industry classifications:
The upper left-hand box in figure 3-1 shows the upstream industries, the upper right-hand box represents midstream industries, and the box below them represents downstream industries. Raw materials are extracted or grown by upstream industries. The midstream companies transform those goods into intermediate goods. Downstream industries purchase the intermediate goods and produce end-use products. An example of this process is the making of an automobile. Upstream producers extract iron ore, copper, oil, and other raw commodities. Intermediate goods producers convert iron ore into steel, and fabricated metal shops transform this steel into the various parts of an auto body. Smelters take the copper ore to produce copper for use by wire manufacturers; these then turn the copper into fine copper wire for the electrical components of automotive electronics. Utilities convert oil or coal into electricity. Auto manufacturers purchase the metal, copper, and plastic parts as well as engines, transmissions, and tires. Using electricity purchased from a utility, the auto manufacturer assembles the various parts into motor vehicles. The process described above is linear and sequential. However, as shown in figure 3-1, the process includes feedbacks from the downstream to the midstream and upstream industries as well as from midstream to upstream industries. For example, farmers purchase transportation equipment such as tractors and trucks from downstream industries, and intermediate goods are consumed by upstream industries. An example of an ATP project enabling a better intermediate good to be used by a raw material producer is found in Pelsoci (2004). In that ATP project, a materials producer created an improved oil drilling part, known as a riser, which allows oil extractors to drill for oil in water depths previously unavailable to them. Another ATP project example involves a producer of catalysts for the natural gas industry. Natural gas producers increase their yield of usable natural gas through improved removal of nitrogen and other impurities.
The NAICS Assignment ProcessBeginning with their baseline report—i.e., the first BRS survey given to all ATP project participants at the beginning of a project—companies are asked to identify potential commercial applications resulting from the technology developed from their ATP project. In subsequent annual surveys, they are asked whether any new applications emerged and whether existing ones are still applicable. In this research, the authors collected all commercial applications listed by participating companies in the BRS beginning with projects that started in 1999, when the web-based version of the BRS was implemented, up to January 31, 2004. Depending on how long a company's project had progressed, each participant completed between one and four BRS reports. Some projects had already been completed; others had been terminated early. Sometimes a company completed a baseline report, but did not report commercial applications until the project's first anniversary. The raw data set resulted in 1,786 observations of potential applications. Because companies were asked in later surveys whether the application was still viable, there are duplicate observations in this set. We coded all commercial applications even if a project had been terminated or the commercial application was indicated to be nonviable in subsequent annual surveys. After eliminating those observations where there was baseline information but no commercial applications and duplicate commercial applications captured in the annual surveys, 852 unique commercial applications remained, involving 372 unique companies participating in 265 unique projects. We examined each project participant individually in order to code its own-industry, i.e., the primary industry in which the company operates. In making this determination, the three-digit NAICS codes (if reported by the company) provided some guidance; we additionally relied on Compustat information for NAICS codes for public companies and the Hoover's database for private companies. If more than one NAICS or SIC code was listed for the company, we used the one that appeared most closely related to the technology used in the project. If neither Hoover's nor Compustat listed any codes, we conducted further web searches to identify the own-industry code. Through this process, each project participant was assigned a unique own-industry NAICS code. A similar process was followed to determine the most common use-industry of each application. In many cases, defining the industry of the commercial application was more difficult than defining the own-industry. The commercial application titles provided by the companies tend to be short; these are frequently either too broad-based (e.g., "Auto") or vague (e.g., "Drug Discovery"). Second, because these products and services are often completely new commercial lines of business, no specific NAICS code yet exists to describe them. We often used the long description of the project that the company provided at the beginning of a project as a way to better understand the applications being proposed. Companies were also asked whether the commercial application was a new product, service, license, or process. This information helped clarify the potential usage for the product, and was helpful in determining appropriate NAICS codes. Examples of The NAICS Assignment ProcessThis section presents three examples of the NAICS assignment process to better illustrate the methodology and the challenges faced in implementing it.
Note that, when displayed in a table, the process may appear straightforward; however, it took much detective work and analysis to assign NAICS codes to ATP participants (own-industry) and applications (use-industries). Ultimately, one can observe how much variety in expertise is brought to the latter joint venture project. The Pig Improvement Company's expertise is concentrated in pig production, but it may be able to bring some fresh insight to shrimp genetics. An animal feed manufacturer adds expertise to the shrimp food side. The fact is that different producers from various industries bring unique skills and knowledge sets to this project. All of these conditions stimulate the potential for spillovers.
___________________ 2. For information on SIC codes for ATP projects between 1993 and 1997, see Powell and Lellock (2000). 3. I nstead of NAICS codes, companies were instead asked to classify their technology in terms of National Institute of Standards and Technology's 7 focus and 43 sub-focus areas. These focus and sub-focus areas are different than the five technology categories ATP uses to classify its projects. Return to Table of Contents or go to next section. Date created: May 25, 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 |