The Case for Doubling US R&D Spending

“We massively underinvest in science and innovation, with implications for our standards of living, health, national competitiveness, and capacity to respond to crisis.” Benjamin F. Jones makes the case in “Science and Innovation: The Under-Fueled E/ngine of Prosperity.” It’s one of eight essays appearing in an e-book on Rebuilding the Post-Pandemic Economy (Aspen Economic Strategy Group, November 2021). Jones offers some vivid reminders that increases in standard of living–not just purely economic, but also other aspects like health–are closely related to investments in new technologies.

Real income per-capita in the United States is 18 times larger today than it was in 1870 (Jones 2016). These gains follow from massive increases in productivity. For example, U.S. corn farmers produce 12 times the farm output per hour since just 1950 (Fuglie et al. 2007; USDA 2020). Better biology (seeds, genetic engineering), chemistry (fertilizers, pesticides), and machinery (tractors, combine harvesters) have revolutionized agricultural productivity (Alston and Pardey 2021), to the point that in 2018 a single combine harvester, operating on a farm in Illinois, harvested 3.5 million pounds of corn in just 12 hours (CLASS, n.d.). In 1850, it took five months in a covered wagon to travel west from Missouri to Oregon and California, but today it can be done in five hours—traveling seven miles up in the sky. Today, people carry smartphones that are computationally more powerful than a 1980s-era Cray II supercomputer, allowing an array of previously hard-to-imagine things—such as conducting a video call with distant family members while riding in the back of a car that was hailed using GPS satellites overhead.

Improvements in health are also striking: Life expectancy has increased by 35 years since the late 19th century, when about one in five children born did not reach their first birthday (Murphy and Topel 2000). Back then, typhoid, cholera, and other diseases ran rampant, Louis Pasteur had just formulated the germ theory of disease, which struggled to gain acceptance, and antibiotics did not exist. In the 1880s, even for those who managed to reach age 10, U.S. life expectancy was just age 48 (Costa 2015). Overall, when examining health and longevity, real income, or the rising productivity in agriculture, transportation, manufacturing, and other sectors of the economy, the central roles of scientific and technological progress are readily apparent and repeatedly affirmed (Mokyr 1990; Solow 1956; Cutler et al. 2006; Alston and Pardey 2021; Waldfogel 2021).

Jones emphasizes some other gains from technology as well. For example, technology can offer flexibility in confronting various threats. Without decades of earlier research, COVID vaccines could not have been developed in less than a year after the pandemic hit: “Whether facing a pandemic, climate change, cybersecurity threats, outright conflict, or other challenges, a robust capacity to innovate—and to do so quickly—appears central to national security and national resilience.”

Moreover, it’s worth remembering that many countries in the rest of the world have active research and development efforts in many areas. The technology frontier is a moving target. The US will either stay near the lead in many of these areas, or fall behind.

In the mid-1990s, the United States was in the top five of countries globally in both total R&D expenditure as a share of GDP and public R&D expenditure as a share of GDP (Hourihan 2020). Today, the United States ranks 10th and 14th in these metrics, and U.S. public expenditure on R&D as a share of GDP is now at the lowest level in nearly 70 years. … By contrast, China has massively increased its science and innovation investments in pursuit of leading the world economically and strengthening its hand in global affairs. China’s R&D expenditure has grown 16% annually since the year 2000, compared to 3% annually in the United States. If China implements its current five-year plan, it will soon exceed the United States in total R&D expenditure.

Jones’s essay reviews the argument, fairly standard among economists, that a pure free market will tend to underinvest in new technologies, because in a pure free market the innovator will not capture the full value of an innovation. Indeed, if firms face a situation where unsuccessful attempts at innovation just lose money, while successful innovations are readily copied by others, or the underlying ideas of the innovation just lead to related breakthroughs for others, then the incentives to innovate can become rather thin, indeed. This is the economic rationale for government policies to support research and development: direct support of basic research (where the commercial applications can be quite unclear), protection of intellectual property like patents and trade secrets, tax breaks for companies that spend money on R&D, and so on.

A key insight is that many innovations build on other insights in unexpected ways. Here are a couple of vivid examples from Jones: the link from Albert Einstein to Uber, and the link from life in hot springs to genetic science.

Uber is a novel business model that has disrupted the transportation sector, and to the user Uber might appear as a simple mobile app enabling a new business idea. But Uber relies on a string of prior scientific achievements. Among them is GPS technology, embedded in the smartphone and in satellites overhead, which allows the driver and rider to match and meet. The GPS system in turn works by comparing extremely accurate time signals from atomic clocks on the satellites. But because the satellites are moving at high velocity compared to app users and experience less gravity, time is ticking at a different speed on the satellites, according to Einstein’s mind-bending theories of special and general relativity. In practice, the atomic clocks are adjusted according to Einstein’s equations, before the satellite is launched, to account exactly for these relativistic effects. Without these corrections, the system would not work. There is thus a series of intertemporal spillovers from Einstein to the GPS system to the smartphone to Uber (not to mention all the other innovations, mobile applications, and new businesses that rely on GPS technology) …

To study DNA, it must first be replicated into measurable quantities, and this replication process depends on many prior scientific advances. One critical if unexpected advance occurred in 1969, when two University of Indiana biologists, Thomas Brock and Hudson Freeze, were exploring hot springs in Yellowstone National Park. Brock and Freeze were asking a simple question: can life exist in such hot environments? They discovered a bacterium that not only survived but thrived—a so-called extremophile organism—which they named Thermus aquaticus. Like Einstein’s work on relativity, this type of scientific inquiry was motivated by a desire for a deeper understanding of nature, and it had no obvious or immediate application. However, in the 1980s, Kary Mullis at the Cetus Corporation was searching for an enzyme that could efficiently replicate human DNA. Such replication faces a deep challenge: it needs to be conducted at high heat, where the DNA unwinds and can be copied, but at high heat replication enzymes do not hold together. Mullis, in a Eureka moment, recalled the story of Thermus aquaticus, knowing that this little bacterium must be able to replicate its DNA at high heat given its environment. And indeed, Thermus aquaticus turned out to provide what was needed. Its replication enzyme was declared by Science Magazine to be the “molecule of the year” in 1989. Mullis would be awarded a Nobel Prize soon after, and the biotechnology industry would boom, opening new chapters of human progress.

When the spin-off effects to other discoveries and inventions are taken into account, the gains to research and development are enormous. What would you have been willing to pay for a COVID vaccine in early 2020? Jones says it this way:

Notably, these social returns are not just good: They are enormous. Effectively, the science and innovation system is akin to having a machine where society can put in $1 and get back $5 or more. If any business or household had such a machine, they would use it all the time. But this machine is society’s. The gains from investment largely accrue to others—not so much to the specific person who puts the dollar into the machine.

Of course, it’s impossible to know in advance exactly what ideas are going to be important, or what firms are going to be success stories. Indeed, one problem with relying on the private sector for R&D is that there is a tendency for firms to focus on the technologies that look most profitable in the short- or medium-terms, rather than building up a broad portfolio of knowledge that can be used in many ways. It’s important to let a thousand flowers blossom–because, if I can mix my metaphors, one of those flowers will grow into a mighty redwood. Jones says it more neatly: “In many ways, the vision of science and innovation needs to be the opposite of `picking winners.’ Rather, we need to `pick portfolios,’ with an emphasis on both increasing the scale of funding and human capital, and the diversity of approaches that are taken.”

Jones has lots of other points to make about technology and research–for example, although it’s widely believed that innovations are more likely to come from younger researchers, this does not actually seem to be true. But the bottom line is that when economists try to calculate the broad social returns to investing in research and development, it’s common to find estimates of annual returns in range of 40-50%. He argues that “a sustained doubling of all forms of R&D expenditure in the U.S. economy could raise U.S. productivity and real per-capita income growth rates by an additional 0.5 percentage points per year over a long time horizon.” And course, these economic gains don’t include the gains to health, or a greater ability to respond in crises, or the benefits of maintaining US global competitiveness.

Jones is also thoughtful in noting that national efforts at research and technology, and at applying those innovations in the broader economy, are not just a matter of budgetary appropriations. It’s necessary to expand the number of researchers and laboratories, which in turn means increasing the pipeline of people with the interests and skills to do that work, which in turn means reaching back to college and high school and elementary school–because someone who, say, leaves fourth grade without being able to do basic arithmetic is likely to have a much harder time becoming a researcher someday. This literature sometimes discusses the problem of “lost Einsteins”–those American children who never got the support and encouragement to develop their underlying abilities in math, science, and innovation.

Another part of the picture–and a faster way to expand US R&D than expanding the pool of students interested in these areas–is to encourage skilled immigration.

In a systematic study of inventors in the United States, Bernstein et al. (2019) examine the role of immigrants in U.S. invention. The central finding is that immigrants are especially productive in inventive activity. Not only do immigrants patent more often than U.S.-born individuals, but their patents are both more impactful for future invention and have greater market value. Overall, immigrants produce twice as many patents as one would expect from their population share. This is consistent more broadly with the STEM orientation of the immigrant workforce. While immigrants make up about 14% of the U.S. workforce, they account for 29% of the college-educated science and engineering workforce and 52% of science and engineering doctorates (Kerr and Kerr 2020). Overall, immigrants have accounted for about 30% of U.S. inventive activity since 1976 (Bernstein et al. 2019).

A similar picture emerges when examining entrepreneurship. Azoulay et al. (2021) study every new venture in the United States founded from 2007 through 2014 and examine whether the founders were born in the United States or abroad. They find that immigrants are 80% more likely to start a company than U.S.-born individuals. Moreover, immigrant founders are more likely to start companies of every size, including the highest-growth and most successful new businesses (see Figure 6).16 Indeed, looking at Fortune 500 firms today and tracing them back to their founding roots, one similarly finds a disproportionate role of immigrant founders—from Alexander Graham Bell to Sergey Brin to Elon Musk. A remarkable finding here is that immigrant-founded firms employ more people in total than there are immigrants in the U.S. workforce.