Taking Account

BEA study derives industry R&D productivity. The chal­ industry level. Not surprisingly, R&D price indexes lenge is to identify the growth the use of a common deflator for Research and development rate of this R&D productivity R&D investment implies more (R&D) and other intangibles are using data available on an ongo­ rapid R&D investment growth widely understood to make sig­ ing basis. for industries that have experi­ nificant contributions to eco-The experimental indexes are enced relatively rapid increases nomic output and growth. based on industry cost weights in input costs, compared with an Estimating R&D for national from National Science Founda­ industry-specific deflator. economic accounting purposes tion (NSF) R&D expenditure For use in deflating R&D for calls for R&D price indexes suit-data, wages from the Bureau of the national accounts, the au-able to deflate nominal measures Labor Statistics (BLS), and inter­ thors concluded that at the of R&D output. mediate input data from BEA. macro level, the impact on R&D A recent paper by Bureau of These cost weights encompass investment and GDP is not Economic Analysis (BEA) econ­ wages for scientists and engi­ greatly affected by the choice of omists Carol Robbins, Olympia neers, wages for support person-industry-specific or aggregate Belay, Matthew Donahoe, and nel, materials and supplies costs, deflators. Given the tradeoff be-Jennifer Lee describes the con­ current-cost depreciation, and tween computational complex­ struction of experimental indus­ other R&D costs. ity and precision, an aggregate try-specific R&D price indexes To adjust for the unobserved deflator for business R&D has to deflate nominal R&D output productivity in the knowledge-substantial appeal. and investment in the absence of creation process, the authors use However, the impact of market prices and quantity mea-BLS multifactor productivity for choosing industry-specific or sures. the nonfarm business sector. aggregate deflators is more pro-Indexes were constructed for The study also compared the nounced at the industry level. pharmaceutical manufacturing, authors' data-intensive ap-Because the authors used a sin-semiconductor manufacturing, proach that estimates R&D costs gle productivity adjustment for motor vehicle manufacturing, separately for the five specific in-all industries, the variation computer system design and re­ dustries with an approach that across industries in the price in­ lated services, and scientific uses a common deflator. dexes is a function of input cost R&D services. The authors calculated that variation. Building from a standard in-the weighted average of produc-This variation in input costs put cost approach recom­ tivity-adjusted input cost …