February 14, 2024
The roots of US innovation clusters
Daniel Gross and Bhaven Sampat discuss the impact of one of the largest shocks to federal R&D funding in US economic history.
A technician changes a bad tube on the ENIAC, the first general-purpose digital computer, built with government funding.
Source: M. Weik, Public domain
Before Silicon Valley became a byword for innovation, Route 128, outside of Boston, was America's technology highway, connecting the country’s premier technology companies and research facilities. However, this first American high-tech cluster likely would not have developed as it did without one of the biggest shocks to federal R&D funding in US economic history.
In a paper in the American Economic Review, authors Daniel P. Gross and Bhaven N. Sampat explain how a World War II research effort jump-started innovation hubs like Route 128 across the United States. Gross and Sampat recently spoke with Tyler Smith about the history of R&D funding in the United States, and the lessons policymakers can take from it.
The edited highlights of that conversation are below, and the full interview can be heard using the podcast player.
Tyler Smith: What was the Office of Scientific Research and Development (OSRD) and how does it fit in within the broader history of the US approach to supporting innovation?
Daniel Gross: The OSRD was created in 1941 as an extension of the National Defense Research Committee (NDRC). The NDRC was authorized by President Roosevelt in June 1940 to harness civilian science to attack military technology problems presented by the growing recognition among policymakers that the United States would be drawn into the war eventually. It was around this time that the United States began investing in research to support progress in military innovation, in particular, because there was a concern that the US military significantly lagged the then-modern technological frontier of warfare.
Bhaven Sampat: In addition to funding, the OSRD also played an important coordinating role. During the interwar era, it was well known that the side with the better science and technology would have an upper hand. And that's both good news and bad news. It was good news because there are pockets of scientific and technological capabilities scattered across the United States. But it was bad news because before the OSRD was created, there was no formal science and technology policy. There were no formal mechanisms to try to coordinate and network the activities of the various actors in the US innovation system. So, in addition to being the first major funding of research outside of agriculture, it also has this important coordinating, networking, portfolio overview and setting approach.
Gross: I think it's also useful to think about what the landscape looked like at the dawn of World War II in the United States. Different sectors of the economy that did research were fairly siloed working on industrial problems. The military had intramural research, for example, the Naval Research Laboratory, that was working on military technology matters, including things that developed rapidly during the war, like remote detection and radar. But a big part of what the OSRD did was play a key role in coordinating priorities, identifying military problems that need solving by actually working with military liaisons, putting those problems under the microscope, directing resources to them, identifying researchers and organizations that could work on them, building ties and networks among different research performers, and ultimately playing a role in getting the results of that research into production and actually into practice.
Smith: Which impacts of the OSRD funding were you the most interested in studying?
Gross: To me, the most interesting outcome is actually the development of innovation capacity in the United States. I think that's really what most of the paper is focused on. We particularly examine regional innovation capacity, looking at the development of technology clusters and evidence that the war triggered the growth of several technology clusters. But beyond that, the bigger thing that I think I'm drawn to understanding better is the capabilities that we see in the US innovation system as well as the direction that innovation took after the war—postwar era communications, electronics, and ultimately the development of electronic computing. Where did all this come from? What explains the strength of certain regions in the modern US innovation system? These are the kinds of questions that these data and this major historical shock and natural experiment are able to provide us some insight into.
Smith: There are numerous characteristics that might affect how desirable a region is for starting an innovative project. How did you isolate the impact of the OSRD funding?
Gross: A lot of what OSRD was funding and working on was pretty new for its time. There wasn't a deep well of existing capacity in a number of these subjects. Microwave radar, for example, was a new technology, and while there was some existing expertise, especially in the industrial sector in radio engineering, that’s not really where the work took place. The major work in that area took place under the direction of a new organization that was created in Cambridge, Massachusetts, which despite its being a major nexus of innovation today, didn't stand out quite as much before World War II as it did afterwards. I would also point out that the parts of the country that had more existing infrastructure assets and specialized knowledge that OSRD could draw on were more innovative before the war than other parts of the country, but they weren't on differential trends. What we do in our work is difference out those existing gaps in levels and look at changes in trends.
These big place-based R&D efforts can, even if they help promote innovation, exacerbate regional inequalities in innovation and patenting that may feed down to jobs, entrepreneurships, and other social outcomes that we care about.
Bhaven Sampat
Smith: I think it might be helpful to illustrate the transformative nature of this funding by looking at a particular county. Do you have any examples that stand out?
Gross: Middlesex County, Massachusetts, is our canonical example. In particular, one of the biggest investments that OSRD made in Middlesex was the radar development program that was centered at the MIT Radiation Laboratory. It was a new organization that was created in 1940 to begin to attack the radar problem. You see that overall total patenting in Middlesex County was higher than other counties in Massachusetts, but pretty stable over the ten or so years prior to the war. You see a big pop, roughly a doubling in annual patenting in Middlesex during and near the end of the war, as the work that was performed there that was funded by OSRD entered into patented inventions. It comes back down, reverting to prewar levels just after the war ends, but then it begins to grow, whereas other Massachusetts counties are generally still pretty stable. You can look at different technologies being developed within Middlesex County, and you'll see similar patterns, particularly in the technology areas which OSRD was supporting—communications and electronic systems. This kind of work comes back to the work that was taking place at the Rad Lab, which is a cornerstone of the story of the Route 128 cluster and of MIT itself.
Smith: Do you think your work sheds any light on what we should expect from new efforts to spur US innovation?
Bhaven: In some sense, OSRD is just a really interesting shock because it's probably the biggest single shock to federal R&D funding in US economic history. But we tried to be very careful at the same time in generalizing from that to the here and now. We spent a lot of time drawing the line between what's very specific to this big bang and what might inform policy today and going forward. I think where we end up is first, this provides evidence that an applied R&D shock—an R&D effort focused on solving specific problems as opposed to just basic science for the sake of funding science and promoting science—can actually have a long-run economic impact. And to that extent, it is consistent with some current policy initiatives focused on mission-oriented or problem-solving R&D—everything from chips and science to ARPA-H. Second, we also show that these big place-based R&D efforts can, even if they help promote innovation, exacerbate regional inequalities in innovation and patenting that may feed down to jobs, entrepreneurships, and other social outcomes that we care about. I think that relates to current-day efforts to think not just about the efficiency of funding, but also equity considerations and spreading out funding geographically across institutions and across individuals.
♦
“America, Jump-Started: World War II R&D and the Takeoff of the US Innovation System” appears in the December 2023 issue of the American Economic Review. Music in the audio is by Podington Bear.