Alan Ladwig, associate administrator for policy and plans at the National Aeronautics and Space Administration, noted that in the agency's efforts to refocus its R&D on cutting-edge science and technology, officials confront the difficulty of measuring the contribution of long-range R&D to the economy and society—a concern compounded by the requirements of the Government Performance and Results Act (GPRA).
The workshop participants identified various types of policy-making needs for innovation information, including data for program development, program evaluation, budget allocations between R&D and non-R&D programs and among fields of research, and the development of standards and regulations. A distinction was made between (1) information acquired through a national aggregate statistical program in the area of science, technology, and innovation and (2) information required for program evaluation, program operation, or program development. The first type of information involves developing industry and national aggregate databases that reflect broad trends in industrial innovation, linked to other economic data, and tracking these developments with relative consistency overtime. The second type of information involves more specialized, often industry- or technology-specific, data. The utility of program-specific data is enhanced, however, if linked to more aggregated data and therefore to an understanding of the broader trends.
Although this workshop focused primarily on aggregate data, it was recognized that as the economy and policies change and as the understanding of innovation processes evolves, the composition of the national innovation data sets must be periodically revisited. At the same time, policy agendas constitute moving targets. The design or improvement of a system of national innovation statistics should focus on establishing a relatively stable foundation or core set of indicators but with the realization that the meaning of particular indicators may change over time. Collectors and users alike should regard indicator data as a dynamic body of information.
At the workshop, several officials of federal agencies with diverse responsibilities described how they perceived their needs for information about industrial innovation. Their comments are summarized in the following.
Tim Brennan, senior economist on the staff of the Council of Economic Advisers, noted that better information on R&D and innovation would inform federal budget decisions regarding the allocation of funds in support of science and technology. To address the basic question of how much the government should spend on R&D and in what areas, information is needed on the benefits and costs of subsidizing R&D and on the market failures that justify government intervention, because private incentives left to themselves will not produce socially optimal results. Markets provide reasonable tests for other activities, but with respect