9
Innovation in the Service Sector and Information Technology
As several industries have expanded relative to the manufacturing sector and information and communications technologies have transformed processes of production and enabled the introduction of new products throughout the economy, pressure to expand the coverage of national science and technology statistics to include these areas has intensified. The NSF's Science Resources Studies Division has extended its R&D survey to nonmanufacturing and is considering new methods of monitoring innovative activity in high-technology and service industries.6 The question is not whether this is desirable but to what extent survey methods and design need to be modified to capture relevant information. A small discussion group was convened to address these issues.
In the discussion, which was led by David Mowery, professor of economics at the University of California Haas School of Business, participants noted that most service and high-technology industries differ significantly from manufacturing in terms of the nature of the output, industrial structure, and competitive dynamics. Mark Myers, senior vice president for corporate research and technology at Xerox Corporation, explained that for several reasons models of manufacturing innovation are of limited use in describing or explaining innovation systems in software and services. First, although patents and trade secrets are important to manufacturers, competitive wars in information technology are waged over architecture. Second, different processes can be competitive in producing for the same product market in manufacturing, but multiple architectures are not sustained in information technology. Third, firms in software and services form innovation systems of complementary relationships, different from those formed in manufacturing systems. The relationships between suppliers and buyers tend to be much closer and more dynamic in services than in manufacturing.
Because the information and communications technologies are undergoing such radical and rapid changes, the process of innovation in these industries is itself a moving target, and there appears to be little agreement on what national statistical programs should track. Many participants remarked upon the need for basic knowledge about the nature of innovation in service industries and the difficulty of defining or measuring service-sector outputs and their quality and quantity. Other research priorities mentioned included
- Assessments of information technology adoption and impacts,
- Assessments of returns to technology development and adoption, and
- Better understanding of the structure and nature of industrial innovation systems in the software and related services industries.
It was noted that one of the best existing sources of information on software is the capitalized development expenses reported in quarterly and 10-K filings with the Security and Exchange Commission, which may enable the development of capital stock estimates for software. The requirement to report capitalized software development expenses is not universally observed, however. To obtain more representative data from this source, compliance needs to be encouraged or enforced.
One participant suggested that a special survey of the use and impact of information technology in different industries could provide critical baseline information and that a model approach might be a recent Statistics Canada survey addressing the use of biotechnology across industries, the impact on product mix, employment, and workforce factors. Others suggested that the NSF form an advisory group to develop an agenda for addressing information technology and service-sector data issues.