US government funding for research and development (R&D) has fallen dramatically since the 1960s, while private R&D investment has tripled. In new research, Arnaud Dyèvre examines the effects of public R&D on firms’ growth and productivity. He finds that public and private R&D are complementary, with public R&D doing the fundamental research, and private firms building on these to create new products and services. He argues that to restore productivity growth in the US, the federal government should end the 60-year-old decline in the share of GDP it dedicates to publicly funded research. 

The history of technological progress in the US is replete with examples of innovations that were first developed with public money. Nuclear fission for civilian use, GPS, the Internet, and even the lighting technology used by the vertical farming industry that was developed by NASA in the 1980s to grow plants in space. But one does not have to look very far to also find examples of some spectacular failures of government interventions, like the government-backed solar panel manufacturer Solyndra going bankrupt in 2011, or instances of the private sector leading the way in the development of transformative technologies, like mRNA vaccines.

For those who support an interventionist government, the decline in the share of GDP dedicated to federally funded R&D in the US over the last 60 years (shown in Figure 1 below) is cause for concern. Public R&D as a share of GDP peaked in 1964 and has been divided by three since then (blue time series). In contrast, private R&D has tripled over the same period (orange time series). This decline in public R&D is significant; even the $96 billions of R&D expenditures enacted by the CHIPS & Science Act, an ambitious bi-partisan bill from the Biden Administration designed to restore America’s technological lead in the production of semiconductors, is a tiny fraction of what would be needed to bring back public R&D spending to its historical high.

Figure 1 – The shift in the funding of R&D in the US, from public to private

Three key differences between public and private R&D

To understand if the consequences of the decline in public R&D in the US are as severe as some claim, we look at historical data on US patents filed over 70 years. By considering the overall history of patented inventions and their sources of funding (public or private), we can go beyond examples and case studies. It turns out that publicly funded patents are very different from private patents, even after controlling for the R&D spent in the creation of these innovations. Of course, patent data is not perfect as not all inventions are patented and public R&D also funds research that is published in scientific papers instead of patents. But patent data is unique in that it covers innovations that have practical applications in everyday life and are thus more likely to be driving productivity.

First, publicly funded patents are much more likely to rely on science. They cite scientific papers at almost four times the rate of private patents. One interpretation of this finding is that public patents are ‘closer’ (in a technological sense) to fundamental knowledge, while private firms are more interested in funding research that has direct commercial applications and are thus more practical.

Second, publicly funded patents appear to be more impactful than private patents, even after accounting for how much money is spent in the creation of the patents, and after controlling for the productivity of the inventors. A patent’s impact is measured by the total number of patent citations it receives, or by the likelihood that it is a breakthrough idea that opens a new field.

Finally, and most importantly, public R&D tends to generate more ‘spillovers’ meaning that their innovations can be used more widely in the economy. This is reflected by the large number of technologies (‘patent classes’) that cite publicly funded patents. For instance, the patent that gets cited by the largest number of patent classes was filed by a private firm in 1989 after having received funding from NASA. It describes a fascinating technology; micro-actuators which are micron-sized mechanical switches that change shape or size when heated. This technology now has countless applications in medicine, consumer electronics and aerospace.

“Veg-03 Ground Harvest” (CC BY-NC 2.0) by NASA APPEL Knowledge Services

Public R&D spillovers boost private firm growth

The larger spillovers of publicly funded innovations have real consequences on the private economy; they increase firm productivity growth; patent production and firms’ own R&D spending. We can show that these effects are causal and not simple correlation. One could be worried, for instance, that firms more exposed to public R&D spillovers are also in sectors that are growing fast. They could also be benefiting from government contracts that increase these firms’ sales. In both cases, a naïve comparison of firms exposed to public spillovers and firms not exposed to public spillovers would yield a spurious, positive association between firm growth and their exposure to spillovers. To make sure these impacts are capturing causal effects and are not mere correlations, we can rely on large and unanticipated R&D funding shocks at the federal level such as the massive boost to space-related R&D by NASA in the mid-1960s (shown in Figure 2 below).

Figure 2 – NASA R&D budget 1950-2020

Note: Inflation-adjusted 2020 dollars

The large shock in the 1960s was triggered by the successful launch of the Sputnik satellite by the USSR, to which the US had to react. The smaller increase in the late 1980s was triggered by the assembly of the Russian MIR space station in orbit (and spurred the research to grow plants in space mentioned earlier). Because these shocks are unanticipated and aren’t linked with the economic conditions of specific sectors in the US, they can be used as random shocks to mimic a Randomised Control Trial, the most reliable way to estimate a causal link in science. Some firms, because of their patent history, will be more exposed to NASA’s research before the funding shocks occur. When the shock happens, these firms are suddenly exposed to a larger flow of knowledge from NASA. We can use such shocks across all federal agencies such as the Department of Defense or the Department of Energy.

This statistical technique generates the following results, which are summarized in Figure 3 below. A one percent increase in patents funded by public money generates no increase in firm size, measured by total sales or employment, as can be seen in the first two-point estimates and confidence intervals. However, it causes a .025 percent increase in firm productivity (‘TFP’), a .024 percent increase in the firms’ own patents count and a .031 percent increase in the firms’ R&D expenditures. In other words, firms are becoming more productive and more innovative, but not bigger.

Figure 3 – Impacts of public R&D spillovers on firm size, productivity and innovation

Importantly, smaller firms benefit more from public R&D spillovers than larger ones. Perhaps because they may not have the necessary resources to conduct the type of fundamental R&D typically funded by the government. 

Some important consequences for the productivity of the US economy 

Do the differences between public and private R&D matter at the national level? Can the decline in public R&D explain the decline in productivity growth in the US? A macroeconomic model of US productivity growth since 1950 can be calibrated with the decline in public R&D shown in Figure 1 and the estimates of impact of Figure 3 to answer these questions. The model reveals that around a third of the decline in productivity growth since 1960 can be explained by the diminishing importance of public R&D.

But this does not necessarily mean that public R&D is ‘better’ than private R&D, it just serves a different purpose. Crucially, public and private R&D are complementary; public R&D does the background work of fundamental research and then, entrepreneurs use these insights to develop products and services that we want to buy. We need a healthy mix of both types of R&D to drive productivity up, but this mix is currently too heavily tilted toward private R&D in the US.

• This article is based on the paper, Public R&D Spillovers and productivity Growth (latest version: June 2024) 
• Please read our comments policy before commenting. 
• Note: This article gives the views of the author, and not the position of USAPP – American Politics and Policy, nor the London School of Economics. 
• Shortened URL for this post: https://wp.me/p3I2YF-ehK