Abstract: Networks and Innovation: Evidence from African-American Patentees, 1821-2004
This paper introduces a new time-series data set that permits an empirical investigation of the effects of networks on innovation. The data have been constructed from a process of matching African American inventors to U.S. patent data between 1821 and 2000. Exploratory research on this data set reveals three periods during which growth in patenting activity among African Americans relative to the overall patenting population significantly diverges: 1890-1899 (increase), 1900-1919 (decline), and 1983-2000 (decline). The preliminary evidence suggests that, over time, increased levels of innovative activity are associated with increases in the ability to form or participate in three distinct networks: patent, research, and financial and entrepreneurial. Conversely, periods during which these networks are undeveloped or underdeveloped are contemporaneous with lower patenting rates. For example, I find that for most of the 20th century African American inventors have been indepaendent inventors, which is contrary to the trend in the overall population of U.S. patentholders, who are increasingly working in larger patent teams. The sample further shows that African American patentees have a disproportionate number of patents assigned to government entities relative to the overall patent-holding population. These results suggest that networks, as studied in business history, may change the direction, level, and quality of innovative activity. The economic significance of this finding implies that, then and now, network activity may result in nontrivial and persistent declines in inventive activity, the source of improvements in technology that allow sustained increases in per capita income.