Work in Progress

The origins of medical and technological innovations often trace back to academic research, yet the channels through which university discoveries reach the corporate sector remain poorly understood. This paper examines how university research affects local corporate innovation by exploiting the dramatic expansion of National Institutes of Health (NIH) funding from the mid-1990s to 2003—which rose from $24.1 billion to $44.96 billion in constant 2021 dollars—as an exogenous shock that differentially affected research activity across commuting zones. Using difference-in-difference event study models that compare innovation across 65 commuting zones containing R1 universities based on their historical NIH funding exposure (1985–1990), we find that a 1 percent increase in pre-period funding generated $1 to $1.5 million in additional yearly funding from 2002 onward, 2 to 3 additional research projects, and 8 to 15 additional NIH-supported publications. This research expansion spurred local innovation concentrated in corporate rather than university patent filing, with particularly strong effects in the biomedical and physics fields.  By 2002, corporate biomedical yearly patent filings increased by approximately 0.2 percent for each 1 percent increase in exposure to pre-doubling NIH funding. Former NIH principal investigators contribute 0.15 new biomedical patent filings per 1 percent increase in pre-doubling funding, while their non-PI research collaborators separately contribute an additional 0.4 filings across all patent categories by 2012.  These direct and indirect PI contributions, while statistically significant, account for only a modest share of the overall rise in local corporate innovation, suggesting that knowledge spillovers mostly operate through channels beyond the funded researchers themselves.

How and to what extent military research fuels civilian innovation? Defense technologies are developed for strategic and security purposes, are frequently classified, and are subject to export-control regimes such as ITAR that explicitly seek to limit diffusion. Yet a large number of emblematic cases suggest that military research has repeatedly spilled over into civilian technologies. This paper provides systematic evidence on these spillovers by exploiting the sharp increase in U.S. Department of Defense (DoD) research spending that followed the September 11, 2001 attacks. I assemble a new dataset that identifies the principal place of performance for roughly 90 percent of all DoD research contracts between 1996 and 2020, including the pre-2008 period when locations were not fully disclosed, and prior work relied on imputation. Using this dataset, I implement a difference-in-differences design that compares local labor markets with high versus low pre-9/11 exposure to defense research. More exposed areas experience substantially larger increases in per-worker defense R&D spending after 9/11, and these shocks translate into higher civilian patenting. The effects are especially strong for electronics-related technologies that are technologically close to defense applications. These findings show that, despite institutional barriers to diffusion, military research can meaningfully stimulate local civilian innovation in related fields.  Draft here