R&D Price Indexes and Real R&D Expenditures in the United States

The rate of inflation in R&D is very difficult to measure. The official government statistics in the United States use the GNP deflator to deflate R&D expenditures. Little is known about the extent to which price indexes for R and D inputs, if they were constructed in various industries, would differ from the GNP deflator. This paper reports the findings of a study that attempts to fill this important gap. Laspeyres price indexes are presented both for R and D inputs and for inputs used in other stages of the innovation process. Also, price indexes for R and D inputs are constructed, assuming that the relevant production function is Cobb-Douglas. On the basis of the results, we estimate the extent and direction of the errors in the official R&D statistics due to the use of the GNP deflator. Our findings, based on detailed data obtained from firms accounting for about one-ninth of all company-financed R&D in the U.S., indicate that the GNP deflator underestimated the rate of price increase for R&D inputs during 1969-79 in practically every industry we studied. The bulk of the apparent increase in real R&D in these industries (which was relatively modest in any event) seems to have been due to the inadequacies of the GNP deflator. One important moral seems to be that more attention and resources should be devoted to the construction of better price indexes for R&D