Conference Agenda

U.S. firms have reduced their investments in scientific research compared to product

development. We use Census data to study how the composition of R&D responds to an increase in the cost of funds. Companies forced to refinance during the 2008 financial crisis made substantial cuts to R&D. These reductions were highly concentrated in basic and applied research, and their impact appears in citation-weighted patent output after three years. We explore several mechanisms and conclude that the overall pattern of results is consistent with an important role for technological competition in determining the composition of firms’ R&D investments

Hiring restrictions for high-skilled foreign nationals hinder domestic banks' production of cutting-edge innovation. We document this fact using the Employ American Workers Act (EAWA), which banned US financial institutions participating in the Troubled Asset Relief Program (TARP) from hiring new high-skilled foreign nationals until the full repayment of TARP funding. We exploit the differential pre-crisis exposure of similarly troubled TARP institutions to the unforeseen EAWA ban to show that the ban not only impeded the hiring of new foreign talent but also led to a reduction in both the quantity and quality of patent filings, including new products and processes in cutting-edge fields such as FinTech, cybersecurity, and payment systems. Banks more exposed to EAWA did not substitute their internal innovation with the acquisition of external patents. In terms of labor market implications, rather than replacing new high-skilled foreign nationals with domestic employees---the stated goal of EAWA---banks paid higher wage premia to retain pre-crisis foreign hires than the prevailing wages of US workers in the same occupations and locations.

We document that artificial intelligence accounts for a significant and growing share of aggregate innovation produced during the past three decades, and is now diffuse across industries and technology fields. We then study publicly traded firms, finding that firms direct their production of innovation toward AI, motivated by their own, and their customers', labor's exposure to AI technology. We interact exogenously measured innovation capacity and AI exposure to instrument actual AI production. Our central findings are that producing AI increases a firm's future stock returns, supported by both higher profitability and lower risk. The results suggest that AI production increases firm value.