Technology flows matrix estimation revisited

During the early 1980s I estimated a highly disaggregated matrix of technology flows from U.S. industries that performed research and development (R&D) to industries expected to use the R&D outcomes. The results, extended to analyze how technology flows affected productivity growth in the 1960s and 1970s, are reported in Scherer (1982a, 1982b, and 1984). In this paper I return to the scene of the crime two decades later to see whether the desired matrix of technology flows could have been obtained using publicly available information, or information that could be gleaned as a by-product of existing surveys, without a costly effort extracting micro-data from a large sample of individual invention patents (introduction).