"Targeted R&D Policies and Innovation Consequences: Evidence from the National Quantum Initiative"

Paper

Policymakers often subsidize R&D in particular domains for social, environmental or national security reasons. Examples include neurodegenerative diseases, solar energy, and quantum computing. Yet little work has explored the effects of these policies on inventive activity outside the targeted domain. There is reason to believe these effects may be substantial. Specifically, invention requires trained and experienced people, and training and experience take time to develop. When the government infuses financial resources into a targeted domain, it cannot instantly create more trained and experienced people in that domain. Instead, the financial infusion incentivizes people to move from other related research areas into the targeted area. As a result, targeting one area may lead to an exodus of talent from related areas, harming innovation in these related areas.

In this paper, we explore this possibility through an investigation of the U.S. National Quantum Initiative (NQI) Act, 2018 legislation that allocated $1.2 billion to accelerate innovation in quantum computing. We find that quantum computing patent applications surged almost 13-fold following the NQI Act, and the population of quantum computing inventors expanded at a comparable rate. However, adjacent technological domains such as cryptography experienced a reversal in their growth trajectory, with annual patent growth falling from 10.5 percent pre-policy to negative 6.5 percent post-policy. Tracing individual inventors, we attribute these declines to the reallocation of high-productivity researchers from adjacent domains.

We also investigate which organizations lobbied for the NQI Act. We find that lobbying organizations, especially, experienced lopsided substitution, losing approximately 39 patents in other domains for every quantum patent gained. Conversely, and surprisingly, organizations that did not lobby expanded quantum patenting while increasing output in adjacent domains. Altogether, our study suggests that efforts to understand legislative interventions and technological progress must look beyond focal technologies and account for labor market dynamics and lobbying.