Latent innovation in local economies
提出基于产业间买卖关系和空间邻近性的地方经济潜在创新度量，该度量捕捉了专利、研发支出等传统统计未反映的创新，并发现其对美国地区收入和就业增长有显著解释力。
We propose a measure of latent innovation in local economies based on spillovers among industries in terms of inter-industry sales and purchases as well as spatial proximity. The proposed measure captures innovation that is not reflected in typical NSF-based statistics such as patents, R&D spending, or science and engineering workers. When the results are mapped for the U.S., regions that one would expect to be highly innovative also show up as such. To determine whether this measure helps to explain economic growth beyond traditional factors such as human capital and agglomeration, and conventional measures of innovation, we estimate simple regression models with income and employment growth as dependent variables. The proposed innovation measure is statistically significant even after we control for rival causes of growth. We suggest that our measure is preferable to conventional innovation indicators for understanding where in the U.S. innovation, more broadly defined, is occurring.