Terence Kealey
President Donald Trump’s cuts to the federal science budgets have provoked vast alarm, yet the cuts are justified. Unfortunately, though, Trump has not justified them, at least not properly. This blog post will do so by rebutting three myths of government science funding: 1) the supposed economic benefits, 2) the supposed health benefits, and 3) the supposed technological benefits. There is little evidence to justify the claims of big benefits from government funding of science.
The Economic Benefit of Federally Funded Science
The federal government started to fund significant science only after 1940. Before 1940, it had funded only so-called “mission research.” Under mission research, an agency will identify a mission (the Coast Guard might identify a need for greater safety at sea, say) and then research the best way of delivering on that mission.
The federal government’s expenditure as late as 1940 on research was, therefore, small, amounting to only 23 percent of US R&D (and only some 10 percent of US pure science) because it was limited to discrete missions. The two largest missions were defense (which was, of course, legitimate) and agriculture (which was less legitimate, because its problem was over-production, which impoverished farmers; trying to raise agricultural productivity would not solve that problem.)
After 1940, defense research expanded vastly, as it had in 1861–65 and 1915–1919. However, the research agencies the federal government had created during the Civil War and WW1 were defunded on the resumption of peace, and after 1945, the federal government embarked on its usual peacetime research contraction. For example, the Office for Scientific Research and Development (OSRD), which had been established in 1941 and funded the Manhattan Project, MIT’s Radiation Laboratory, and other vast WW2 research missions, was shuttered in 1947.
However, the post-war trend to federal research retrenchment was reversed in 1950 because of the threat of the Cold War. So, in 1950, the National Science Foundation (NSF) was established, and the National Institute of Health (NIH) was expanded. Indeed, the federal funding of research expanded dramatically, peaking in 1964; so, whereas in 1940 the federal and state governments had accounted for only 23 percent of US R&D, by 1964 it was 67 percent; and whereas in 1940 the federal government had accounted for only 10 percent of US basic science, by 1964 it was 75 percent.
In the space of a quarter of a century (1940–1964), therefore, US research had transitioned from laissez-faire to near-nationalization, and where a company seeking innovation had once raised its own research funds, by 1964 it was writing a grant application to the Department of Commerce. Where a university seeking innovation had raised its own research funds, by 1964 it was writing grant applications to the NSF and NIH. This vast expansion saw, however, no stimulation of US GDP growth.
The creation of the NSF, though, represented more than just the creation of another federal science agency: it represented a shift in doctrine. In response to Vannevar Bush’s 1945 book Science, the Endless Frontier, the NSF was created to fund not mission research but, rather, basic or pure research. Bush—a civilian scientist from MIT and the Carnegie Institution—had led the OSRD during the war, and dreamt of leading a similar body in peacetime. So he wrote Science, the Endless Frontier to argue (i) that pure or basic science was the basis of advances in technology, but (ii) industry would always neglect pure or basic science. Bush, in short, advocated the linear or pipeline model of research:
government-funded pure science → technology → which industry turns into wealth
The advantage of that model was that it would provide Bush with another vast federally funded science agency to run, even in peacetime. The problem with that model, though, is that it is a myth: the government funding of pure science does not generate economic growth. So, under regimes of research laissez-faire, the UK had led the world through the first Industrial Revolution, and the US had led the world through the second Industrial Revolution. France and Germany, on the other hand, had subscribed to the linear model; their governments had funded pure science capaciously, yet they had failed even to converge on UK GDP per capita, let alone (as the US had) overtake it, circa 1890.
The linear model fails for the reasons Donald Stokes of Princeton University described in his 1997 book Pasteur’s Quadrant: Basic Science and Innovation, which can be summarized as:
Almost all new technology builds on established technology, not basic science, and industrial technological research is largely mission research.
Inasmuch as industry needs basic science, industrially funded and privately funded R&D generate all the basic science an economy needs.
Underpinning the myth of the linear model have been two well-known economic papers by Richard Nelson and Kenneth Arrow. In these papers, Nelson and Arrow made two major arguments.
First, they supposed that knowledge is explicit and so spills over freely. Economists of economic growth continue to model knowledge as explicit: not one of the four standard economists’ books on economic growth describes knowledge as tacit. Each treats it as explicit. Consequently, they claim that companies are disincentivized from investing in R&D for fear that their discoveries will spill over, to be exploited by their competitors.
Actually, we have known for a millennium that knowledge is tacit, so it spills over only to other entities doing similar research. Since companies actively seek to copy from each other and cannot copy unless they are also active researchers, it is the tacticity of knowledge that incentivizes the private funding of science.
Nelson’s and Arrow’s second argument was that governments should fund science because companies’ proprietary knowledge would render the market imperfect. To extrapolate arguments from welfare economics into innovation economics (where markets are quite properly modeled as imperfect) is a category error about which Schumpeter had warned in his 1942 book Capitalism, Socialism, and Democracy:
In capitalist reality as distinguished from its textbook [neoclassical] picture, it is not that kind of [price] competition that counts but the competition from the new commodity, the new technology … which commands a decisive cost or quality advantage and which strikes not at the margins of the profits and the outputs of existing firms but at their foundations and their very lives.
Yet Nelson’s and Arrow’s papers have not only been cited over 23,000 times, but they are now being cited more than ever.
Importantly, a series of empirical studies from a succession of government departments, including the Office for Technology Assessment, the Congressional Budget Office (twice), the OECD, and the Bureau of Labor Statistics, have shown that government-funded research has no positive impact on economic growth rates. But these governmental studies have been relatively ignored by the academic literature, which is committed to the paradigm that governments should fund university science, and the futility of trying to discredit theory with facts has discouraged even government departments from continuing to publish such studies, the last of which appeared in 2007.
The Health Benefits of Federally Funded Science
John Early has shown that, during the first half of the 20th century, life expectancy at birth for US citizens grew at a steady rate from circa 47 years (in 1900) to circa 70 years (in 1950); but the rate of increase slowed after 1950, which coincided with the vast post-war increase in NIH funding. Had the earlier rate of increase been maintained, life expectancy at birth for US citizens would today be circa 97 rather than today’s circa 80.
Perhaps the slowing of the rate of increase in life expectancy at birth was unconnected with the rise in the NIH’s funding, and perhaps the rise in the NIH’s funding might have reduced the slowing of the rate of increase in life expectancy at birth. Still, these are not claims even the NIH makes. There is, therefore, no obvious evidence that the vast expansion in NIH funding has improved health in the US.
The Technological Benefit of Federally Funded Science
The post-WW1 era saw the highest rates of Total Factor Productivity (TFP) growth in the US, but the post-WW2 era, which witnessed the vast expansion of the federal government’s funding of research, saw a marked fall in the rate of US TFP growth. There is, therefore, no obvious evidence for the vast expansion in federal research funding having accelerated technology in the US.
The myth is that government funding for research supplements private funding. Actually, it crowds it out. This has been modelled theoretically, but it has also been shown empirically. So, since the 1960s, in acknowledgment that the expansion of government funding of research post-WW2 had not accelerated rates of economic, health or technological growth, the government funding for research has been declining across the western world, and—chronicling the period between 1981 to 2013—two Italian economists, Daniele Archibugi and Andrew Filippetti found that publicly funded R&D amongst the lead OECD countries fell from 0.82 percent to 0.67 percent of GDP. Yet industry-funded R&D rose from 0.96 percent to 1.44 percent of GDP over the same period.
Thus, a fall in publicly funded R&D of 0.15 percent of GDP coincided with a rise in industry-funded R&D of 0.48 percent of GDP. This is a dramatic difference, which suggests that, for every $1, £1 or €1 that a government withheld from R&D after 1981, the private sector put in $3, £3 or €3. It also suggests that, for every $1, £1 or €1 that a government had originally put into R&D, it had crowded out $3, £3 or €3 of private money.
Crowding out has been chronicled anecdotally in ways that anticipate President Trump’s attack on US federally funded research. In 1969, the Office of the Director of Defense Research and Engineering, traumatized by the US military’s failures in Vietnam, published Project Hindsight [AN1] to analyze 700 research ‘events’ that had led to the development of twenty weapons systems. Hindsight found that only two of those 700 research events were in pure science. In other words, the government funding of pure science had not strengthened the US’s defense capabilities.
Dismayed, in 1973 Senator Mike Mansfield (D‑MT, Senate Majority Leader 1961–1977) drove amendments to the Military Authorization Acts to block D/ARPA’s funding of pure science. At the time, Mansfield’s amendment—for making so many pure D/ARPA scientists redundant—was widely denounced for destroying American science. But where did the scientists go? Here is Charlie Gere in his 2002 book Digital Culture:
The first head of XeroxPARC was Bob Taylor [of] ARPA’s computing research arm… The Mansfield amendment and the presence of Taylor at XeroxPARC meant that many talented computer scientists and researchers who had been ARPA-funded were drawn to the Centre [i.e, XeroxPARC].
That is where they invented the personal computer. So, it was the former ARPA researchers who, transplanted to XeroxPARC, invented the mouse, Windows, pop-up menus, and the trash can (indeed, the graphical user interface) as well as the laser printer and the personal computer, including, of course, the Alto. Oh, and they also helped pioneer an Ethernet network and sent things called ‘emails.’
Therefore, because the US’s D/ARPA’s pure scientists were made redundant, XeroxPARC pioneered the personal computer. Without the redundancies, D/ARPA’s advances would have remained tied to the tacit knowledge of the agency itself, and their commercial exploitation would have been frustrated. Without the redundancies, indeed, the Aladdin’s Cave moment when, on seeing inside Xerox PARC, Steve Jobs said “ … within, you know, 10 minutes, it was obvious to me that all computers would work like this someday” would not have happened.
Conclusion
The US federal government launched its vast expansion of research funding in the wake of the launch of Sputnik in 1957, with little to show for it.
President Trump is, therefore, justified in viewing the federal government’s funding of science skeptically. However, two caveats must be made. First, President Trump has not tried to justify his cuts in a scholarly way. Second, he introduced changes abruptly. In 1933, Keynes wrote that such policy changes should be introduced cautiously, to allow adaptation: “Those who seek to disembarrass a country of its entanglements should be very slow and wary. It should not be a matter of tearing up roots but of slowly training a plant to grow in a different direction.”
[AN1]Link