In their paper, “Lockdowns and Innovation: Evidence from the 1918 Flu Pandemic,” Enrico Berkes, Olivier Deschênes, Ruben Gaetani, Lin, and Severen investigate whether government interventions to control the 1918 pandemic affected rates of technological innovation. Borrowing from historical experience, the authors consider possible policy implications of the COVID-19 pandemic on longer-run productivity growth.
The 1918 influenza pandemic was severe, lasting from mid-1918 through April 1919. The second wave that began in September 1918 was responsible for the majority of deaths. Starting in mid-September, U.S. cities adopted numerous nonpharmaceutical interventions (NPIs) to limit the spread of disease by reducing social interactions. These NPIs included bans on public gatherings, school closures, and isolation/quarantine for suspected and known cases. The NPIs varied by city in type and duration.
How might limits on social interactions affect innovation? Extensive research has shown that the degree of social interaction among innovators in cities impacts knowledge transfer channels.1 These “localized knowledge spillovers” are instrumental in the development of new ideas. Cities are more likely to be centers of innovation, and inventions occurring in cities are often complementary. (Inventors in close proximity tend to cite each other’s work more frequently.)2 Knowledge spillovers, to the degree that they involve face-to-face interactions, could conceivably be interrupted under government restrictions to control disease spread. The authors investigated to what extent knowledge transfer channels were diminished during the 1918 pandemic because of NPIs.
The authors examined patenting rates before and after the pandemic, comparing the rates in cities that adopted strict NPIs to those in cities that had less-strict NPIs. They relied on monthly data on NPIs adopted by 50 U.S. cities and patenting rates (the number of patents issued in a city divided by its population). The authors classified cities as having either relatively long NPIs (restrictions lasting more than 90 days) or relatively short NPIs (less than 90 days). The 50 cities in their sample represented just over one-fifth of the U.S. population and almost 40 percent of the patent filings recorded (based on 1910 figures).
The authors’ results show that during the pandemic, cities with long NPIs did not have lower rates of patenting compared to short-NPI cities. Additionally, after the pandemic ended in April 1919, cities with long NPIs saw patenting rates increase by 7 to 12 percent relative to short-NPI cities.3 The authors write, “NPIs during the 1918 pandemic might have reduced patenting in the short run by reducing labor inputs, capital inputs, and idea inputs. But these negative short-run effects may have been limited by the type and short duration of most NPIs. In the medium term, by improving health and reducing mortality, NPIs may have preserved inventor labor supply and organizational capital, increasing patenting rates.”
The authors also find that long NPIs had larger positive effects on patents with multiple inventors and on patents with external assignees (often sponsored by industrial or corporate research and development [R&D] labs).4 The authors write, “These types of patents may reflect more complex invention processes and may also rely more on social interactions.”
Moreover, it was patents in areas of rapid technological expansion at the time, such as electricity and mechanical engineering, that benefited the most from longer NPIs. The authors suggest that patenting in high-growth areas is more prone to technological and market risk. As a result, these areas may have benefited more from long NPIs because cities with longer restrictions likely provided an environment that was more conducive to invention through reduced mortality, dampened uncertainty, and a more stable financial and business environment.
A recent study showed that Prohibition (in place from 1920 to 1933) had negative effects on rates of innovation, which the researcher attributed to reductions in interpersonal communications, especially informal exchanges at drinking establishments.5 To explain the differences between the results of the two studies, the authors of “Lockdowns and Innovation: Evidence from the 1918 Flu Pandemic” suggest that the type of government restriction matters — in the case of the 1918 pandemic, school closures were the most commonly used NPI,6 which likely had less of an impact on knowledge flows important for inventors. The authors also explain that all of the NPIs during the 1918 pandemic, even long NPIs, were relatively short lived, which may indicate “limited long-run effects on patenting through a social interaction channel.”
The authors make comparisons between the 1918 pandemic and the COVID-19 pandemic in terms of its impact on innovation. Both used NPIs, although in 1918, the restrictions were generally shorter and less extensive. During the COVID-19 pandemic, internet communication has likely substituted for some in-person social interactions, helping to promote the flow of ideas. As a result of the differences in the two pandemics, the authors say, it is difficult to make definitive conclusions on whether longer NPIs benefit innovation. Nevertheless, the authors’ evidence indicates that “the decrease in local interactions constitutes only part of the effect of NPIs on invention rates,” and whether longer NPIs positively or negatively impact rates of innovation in a city depends on the relative effects of policy on a range of behavioral and economic responses.
Previous studies on the economic consequences of the 1918 pandemic have focused on contemporaneous measures of economic activity, including the level of employment in the manufacturing sector. In contrast, the authors’ novel approach focuses on the 1918 pandemic’s impact on innovation, and, as a result, they are able to link current activity in 1918 to longer-run productivity growth. Their paper provides useful inputs for (i) researchers working to understand the implications of the COVID-19 pandemic on innovation and, in turn, future productivity growth, and (ii) policymakers as they consider the future use of NPIs to address pandemics and their impact on economic recovery.
- For example, as found in Edward Glaeser, “Learning in Cities,” Journal of Urban Economics, 46 (1999), pp. 254–277.
- Ina Ganguli, Jeffrey Lin, and Nicholas Reynolds, “The Paper Trail of Knowledge Spillovers: Evidence from Patent Interferences,” American Economic Journal: Applied Economics, 12 (2020), pp. 278–302.
- During the pandemic, long-NPI cities had smaller and statistically insignificant increases in patent rates relative to short-NPI cities.
- Tom Nicholas, “The Role of Independent Invention in U.S. Technological Development, 1880–1930,” Journal of Economic History, 70 (2010), pp. 57–82.
- Michael Andrews, “Bar Talk: Informal Social Interactions, Alcohol Prohibition, and Invention,” Working Paper (2019).
- Howard Markel, Harvey B. Lipman, J. Alexander Navarro, et al., “Nonpharmaceutical Interventions Implemented by U.S. Cities During the 1918–1919 Influenza Pandemic,” JAMA, 298 (2007), pp. 644–654.