The big read: is the shift to private R&D holding back growth?

In 2020, an extraordinary threshold was crossed in the world of research and development. 

For the first time since records began, a single company, Amazon, spent more on R&D than the entire German state (see chart below). Germany, birthplace of the modern research university, home to venerable science institutions such as the Max Planck Society, and the biggest public spender on R&D in Europe, is now comfortably outspent by Jeff Bezos’s delivery and web services empire. 

This was not a freak occurrence. The previous year, R&D spending by Amazon, as well as Alphabet, the parent company of Google, both overtook the UK’s public R&D budget. The year before that, France was outspent by Alphabet too. Italian and Spanish R&D budgets have been eclipsed by US tech giants for at least a decade. These figures largely exclude the ongoing spending splurge on data centres to power large language models.

This triumph of US corporate R&D partly reflects the enormous profits that have flowed to the likes of Alphabet and Amazon during two decades of turbocharged digitalisation and, to critics, monopolisation.

But it’s also symbolic of a much broader trend. Over the past 40 years, across the rich world, R&D has increasingly been funded by private companies, rather than the state or public bodies such as universities. Booming spending by companies such as Amazon has masked flat or, in the US, falling public support, which the current US administration wants to cut by a further 35%. 

Until very recently, policymakers and politicians on both sides of the Atlantic have largely shrugged their shoulders at this private shift. They often assumed that it doesn’t matter much who funded R&D, so long as more money flows into the system, and many have welcomed the surge in corporate spending. The EU, Germany and the Netherlands, for example, all have overall R&D spending targets which are agnostic on who stumps up the cash. 

But a new generation of economists say they now have the first empirical evidence that this reliance on privately funded R&D may have been a historic mistake.

This debate is still just beginning. But if they are correct, the failure to invest in public research helps explain why productivity gains due to technological innovation in the rich world have slowed to a crawl, leading to anaemic growth that aids the rise of radical, anti-establishment parties in Europe and the US. 

“There is this growing consensus [. . .] that the decline in publicly funded R&D has contributed to the deceleration in productivity growth,” said Andrew Fieldhouse, an assistant finance professor at Texas A&M University. He’s one of the economists whose work challenges the shift to private R&D, and it has recently caught the eye of some governments. “I think a lot of policymakers are concerned,” he said. 

“It seems like the current funding of R&D is off balance at the moment in the US,” echoed Arnaud Dyèvre, an assistant professor of economics at HEC Paris, and author of another paper that questions the private turn in R&D. 

His work suggests that US R&D spending should trend back towards roughly where it was during the 1960s Space Race, with about two thirds public, and one third private. That would be a huge shift, as business funds 75% of current US R&D. European states should also invest much more in basic research, said Dyèvre. “You do need a healthy mix.” 

It’s not yet completely clear why public R&D spending might be more beneficial than private. But one leading explanation is that it funds more basic research, advancing our underlying understanding of nature and humanity, which has a bigger spillover impact on the economy than the development of new products.

So far, this is a wonkish debate in a niche corner of economics. But it carries huge implications, well beyond research labs. Here, Science|Business looks in detail at the private shift in the R&D system, how it happened and why economists like Fieldhouse and Dyèvre now think it may have been a major error with grave economic consequences. 

The private turn in R&D

The privatisation story of the 1980s is well known: spearheaded by Ronald Reagan in the US and Margaret Thatcher in the UK, a pro-market consensus gripped much of the rich world. 

Everything from British social housing to Austrian airports, French banks and Italy’s motorways operator were controversially sold off, as governments tried to reduce spending and shed loss-making state industries. Between 1984 and 2004, governments across more than 100 countries sold off nearly $1 trillion of state owned enterprises, the OECD has estimated. 

But with far less debate, the R&D system in rich countries also shifted towards the private sector. 

It’s not as though universities and research institutes were sold off to companies, as happened to water companies or railways. Instead, governments simply neglected public R&D budgets, said Andrea Filipetti, a researcher at the National Research Council of Italy, who has argued the economic consequences of this private turn have been “ignored.” Meanwhile, corporate R&D kept growing, leading to a widening gap. 

In some countries, such as France and Italy, this process started from a low base (see graph above). In Germany, the private trend is a gentle, albeit uneven, slope upwards. 

In the US, meanwhile, the shift to private R&D happened much more rapidly. But despite these differences, the trend is visible across the globe. 

However, there’s one important transatlantic difference. While in Europe public R&D spending has merely stagnated as a proportion of GDP, in the US it has dropped dramatically, although some of this decline is down to the winding down of military spending after the end of the Cold War (see graph below). US businesses also accelerated their R&D spending much faster than in Europe, leading to a turbo-charged private takeover in the US system.

Shifts under the hood

Everything else being equal, rising business R&D is still a good thing for the economy, economists stress. No one is criticising companies for spending more; quite the opposite. 

But the problem, some economists say, is that this boom in corporate spending has masked stagnant or falling levels of public support for research. 

“If you look at the United States’ R&D share of GDP, if anything, it’s trending up a little bit, and there’s no immediate cause for concern,” said Fieldhouse. “But if you lift up the hood, there’s a big compositional shift going on down there.” If the funding source of R&D matters, “that compositional shift may be quite problematic.” 

Policymaker indifference

On the whole, policymakers do not look under the hood. They target overall R&D spending, without much attention to the public-private split. 

The EU has a longstanding, and long missed, target for its member states to spend 3% of GDP on R&D, regardless of the source. Of 27 EU countries, only six meet this goal, while seven spend less than 1%. 

European countries also tend to look at overall R&D spending. Over the past decade, Germany set itself a target to spend 3.5.% on R&D, largely achieved by its big-spending carmakers. The UK hit a 2.4% goal in 2022 thanks to a controversial counting change by the country’s statistics office. The Netherlands’ R&D plan, agreed last year, sets a blanket target of 3%.

Meanwhile, Donald Trump’s administration wants to cut federal research spending by 35%, although the US Congress looks likely to blunt these federal spending reductions.  

But even under Joe Biden, US policymakers seemed relatively unconcerned about the country’s shift to private R&D. In 2024, Arati Prabhakar, the then-director of the White House Office of Science and Technology Policy, welcomed a “huge surge” in corporate R&D that meant the US still maintained a spending lead over China, despite flatlining federal spending. 

Patenting Einstein

Economists have long agreed that R&D needs some kind of government encouragement. This idea dates back to at least 1875, when the British statesman Lord Derby told a parliamentary commission that science requires state support because research results “bring absolutely no pecuniary advantage to the person engaged in working them out.” To take an extreme example, there would have been little business reason to employ Albert Einstein, as you can’t patent the general theory of relativity.

It’s easier for companies to profit from more applied, patentable R&D. But even so, the gains of discovery can’t be fully captured by companies. Knowledge leaks out and can be used by rivals, so businesses systematically underinvest in science and its application. 

To solve this underinvestment problem, governments, particularly since World War Two, have doled out research grants and funded universities, given companies tax credits to perform R&D, and granted patents, so that firms reap more of the rewards of invention. The point of these interventions is to boost both public and private spending on R&D.

Public vs private returns

But what has not been empirically researched, until recently, is the relative economic benefits of publicly versus privately funded R&D. In other words, is it better, say, to dish out public grants to university researchers, or incentivise corporations to invent new products?

Fieldhouse, working with Karel Mertens, director of research at the Federal Reserve Bank of Dallas, opened up this research question with a paper published in 2023. Using a huge and meticulously compiled dataset of changes to US federal R&D funding since World War Two, they concluded that public science significantly and enduringly boosted total factor productivity (TFP), an economic measure of technology-driven growth. 

“You see big increases in employment of scientific researchers, STEM [science, technology, engineering and maths] PhD and doctoral students, patents, technology books, all sorts of different measures of productivity,” said Fieldhouse.

Public advantage 

Overall, public R&D has a rate of return of 140-210%, they concluded. Crucially, this is a much higher estimate than the 30-70% rates of return economists calculate for private R&D, said Fieldhouse. These are rough estimates, but suggest a big public advantage.

Dyèvre (right), in a 2024 paper, came to a similar conclusion, this time using data from individual US companies, rather than the economy as a whole. Spillovers from public R&D are three times as impactful on firm productivity as private, he found.

He estimated that US cuts to public R&D account for a third of the decline in TFP growth since 1950. In other words, failure to support public science could be one of the root causes of our current economic predicament.

First empirical work

“These studies are very, very important,” said Benjamin Jones, an economist at Northwestern University who researches innovation and scientific progress. 

Economists have previously theorised that the returns to public science might be extremely high, said Jones. 

There’s also been a surge of economic research in recent years that tracks how public science filters into real-world products, such as how grants from the US National Institutes of Health spur pharmaceutical and biotechnology patents.  

But what’s novel about the two most recent papers is that “it's empirical work that helps colour in the relative returns on private versus public R&D,” said Jones. 

Why are public returns higher?

If Fieldhouse and Dyèvre are right, why might public R&D spending be more economically beneficial than private? 

Their findings, at first glance, sound counter-intuitive. Corporate R&D divisions, after all, have to create saleable products to survive. University-based scientists don’t. “This is the trillion-dollar question,” says Fieldhouse. 

One leading explanation is simple, and will be instantly familiar to most scientists. Public funding is much more likely to back openly published basic research: often curiosity-driven science that tries to understand the universe, rather than develop a specific technology. 

This creates more beneficial spill-overs in the rest of the economy, because anyone can pick up and use this understanding. The germ theory of disease, which revolutionised medicine from the 19th century, is inherently impossible to commercialise, for example. 

Free rider problem?

One objection to spending more on basic science is precisely because it’s open and unpatentable, other countries will benefit just as much as the funding state, creating a free-rider problem.

But the economists who spoke to Science|Business aren’t too worried. They point out that geography and face-to-face connections are still crucial in turning fundamental science into real-world inventions. In Boston, the London-Oxford-Cambridge triangle in the UK, or Munich, say, the tacit knowledge of leading university researchers is an essential part of high-tech cities. 

Basic research creates a “pool of common knowledge” all private firms and broader society can use, as Filipetti puts it. But if basic science gets cut, “this amount of common knowledge shrinks over time.” 

Basic research 

This is far from a new argument. The European Research Council and universities have long made this apparently obvious case for basic research, although without much success. 

And it’s hardly controversial among economists and policymakers that fundamental research needs state support, and fulfils a very different function from commercial R&D.

“Basic research, and to some extent applied research, have a higher spillover potential than experimental development,” says Fernando Galindo-Rueda, a senior economist at the Organisation for Economic Cooperation and Development. “Experimental development” is the OECD’s term for the type of development than counts in R&D statistics. It must generate new knowledge, for example. 

Basic research’s results are typically widely available for others to use, for example through peer-reviewed publications. And because it’s seen as a public good, government funding of R&D in most countries has traditionally tended to support basic and applied research, rather than experimental development, he adds.

But what’s new about the Fieldhouse and Dyèvre findings is that they make the case that the private shift in the R&D system may have tipped this ratio of basic research to technological development out of balance, and that this may have slowed economic growth. 

Related articles

• Czechia re-organises research activities with new innovation-heavy economic strategy
• EU R&D spending increases, but still falls short of 3% target
• How FP10’s budget compares to Horizon Europe

“It's not necessarily that the state is better at funding R&D,” Dyèvre said. “It's mostly that the state has been [. . .] focusing its effort on more fundamental, more basic [research].”

“Maybe it doesn’t matter so much if it’s OpenAI or the National Science Foundation funding basic research,” echoes Fieldhouse. But the point is that private firms are much less likely to back basic research than public funders. 

Bell Labs nostalgia

Of course, some companies have successfully backed groundbreaking fundamental research. The classic example is Bell Labs, the R&D arm of US telecoms firm AT&T, whose researchers won nine Nobel Prizes in physics. More recently, Google’s London-based AI lab Deepmind scooped the chemistry prize in 2024 for the protein structure prediction tool AlphaFold, which was largely made freely available.

That said, firms unsurprisingly tend to focus on D, rather than R. In the US, firms funded 35.9% of the country’s basic research in 2021, according to the most up-to-date statistics, but 87.6% of experimental development. 

There’s lots of nostalgia for the likes of Bell Labs, said Richard Jones, emeritus professor of materials physics and innovation policy at the University of Manchester. However, the evidence is that the last few decades “have seen a real decline in the amount of basic science that gets done in the private sector.”

Disappointing defence R&D

In contrast to public science, private R&D knowledge is much more likely to be protected, either in the form of intellectual property or the “dark matter” of trade secrets, said Jones at Northwestern. “They're not going to spill over in an obvious way.” 

The one field where public R&D had no wider economic benefit, according to Fieldhouse’s work, is defence. This could be because most of this Pentagon spending was in classified weapons development, not basic science, he said, limiting the knowledge spill-over benefits on the rest of the economy. 

The innovator’s dilemma

A lack of basic science isn’t the only reason why private R&D spending might be less beneficial to the economy. Another explanation may lie in the so-called innovator’s dilemma. This is a fairly mainstream idea in economics that big, incumbent companies have less incentive to truly innovate because radical new products might cannibalise their existing sales.  

“Apple will not want to develop a technology that eats into the demand for the iPhone,” said Dyèvre. This means incumbent firms’ R&D spending often goes on making “incremental” improvements to their current offerings, he explained, rather than “disruptive” ideas that lead to whole new product lines and technologies. 

Germany’s deceptive spending?

This opens up a whole new way of thinking about R&D statistics. On paper, Germany appears to be an R&D powerhouse, spending €132 billion in 2023, a healthy 3.1% of its GDP, not far off the US, and one of the highest in Europe. And yet its economy has been sluggish for the best part of a decade.

Dig into German R&D figures, however, and the picture becomes murkier. Around €40 billion of its total spend is from the incumbent German car giants Volkswagen, Mercedes-Benz and BMW, who constitute Europe’s top three corporate R&D spenders. 

But with hugely profitable petrol and diesel car businesses stretching back decades, these carmakers all failed to lead the transition to electric vehicles, and are now playing catch up with Chinese pioneers like BYD.

In 2024, academics made a related argument about US incumbents. For all their R&D spending, giant US tech firms are now stymying innovation by buying off start-ups whose inventions might threaten their business models, they argued. Vast private R&D budgets, in other words, may not guarantee breakthrough invention if spent by entrenched companies. 

Scrap tax credits?

Over the past few decades, governments have ratcheted up tax credits to incentivise businesses to do more R&D, says the OECD’s Galindo-Rueda. The evidence from recent OECD work is that this has boosted experimental development more than research, he went on. 

Since 2000, spending on tax credits in the EU has increased more than fivefold as a proportion of GDP, according to the OECD. The UK, for example, handed out £7.6 billion in tax credits in 2023-24, roughly enough to fund another three Oxford Universities. 

But if Fieldhouse and Dyèvre are correct that public R&D is economically more valuable than private, partly because it backs more basic science, this suggests that governments should think about diverting tax credits to public science. 

This is “economics 101,” said Dyèvre, as tax credits are likely to help companies do applied research, not the higher-return basic science that is impossible to patent. 

“There's uncertainty here, of course, but $1 reallocated from an R&D tax credit to [public] science might be a better investment,” said Jones at Northwestern. But he would much rather see funding for R&D as a whole grow, given how little governments spend on it in general. 

If R&D tax credits are “boosted at the expense of other programmes” this “risks reducing support for applied research in government labs or even business that help connect basic research with both the marketplace and societal goals,” said the OECD’s Galindo-Rueda. 

Why now?

If, as this new research suggests, neglect of public science since the 1980s explains part of the rich world’s economic slowdown, this begs an obvious question: why haven’t economists and governments investigated this before? 

For a start, it’s only relatively recently that economists have had the data and methods to do work like that of Fieldhouse and Dyèvre, said Jones at Northwestern. “The big data revolution [. . .] it’s also come to economics,” he said. 

But there’s also been an “ideological shift” in economics that has led researchers to investigate the returns to public, as opposed to overall, R&D, said Dyèvre. 

That interest was partly piqued a decade ago by The Entrepreneurial State, a book by the University College London economist Mariana Mazzucato, which set out several case studies in which public discoveries fed through into private innovation.

Since then, economists have tried to see if there is quantitative evidence to back up Mazzucato’s anecdotes about state-driven innovation, Dyèvre said. “There's been a flurry of research in this area recently,” said Fieldhouse. 

Are policymakers listening?

For now, Fieldhouse (left) and Dyèvre’s work is part of a niche conversation among research policymakers and economists. But if the conventional wisdom about R&D spending shifts, the implications could be significant. 

The US Congressional Budget Office recently published a new paper estimating returns to public R&D similar to those identified by Fieldhouse. Although the Trump administration wants to cut public R&D, these kinds of behind-the-scenes accounting shifts may well outlast him.  

“I think the consensus among policymakers has changed rapidly over the past few years, in particular in the US,” said Dyèvre. “With the exception of the current administration, a lot of policy folks are convinced more public funding for R&D is needed.” 

Fieldhouse has spoken several times to the British government about the implications of his research, which is now cited in official UK research policy documents that make the case for more public R&D spending. “The Brits are terrified about slowing productivity growth and trying to think through what they can do,” he said. 

More money for companies

However, UK spending plans for the rest of the decade show the country increasing R&D incentives for new companies, while money for “curiosity-driven” research will flatline. 

As for implications for the EU, Dyèvre and Fieldhouse’s work is based on US data, so some caution is needed. And boosting public research is far from a silver bullet, . The bloc also needs to crack other intractable problems, such as integrating its single market. That said, “most countries in Europe would benefit from higher spending in public R&D dedicated to fundamental science,” Dyèvre said. 

But as Europe frets about its immediate economic woes, the focus instead is on corporate innovation. For example, as part of a new economic strategy, Czechia is prioritising applied invention, not basic science. 

And in the EU’s next research and innovation programme, which starts in 2028, the European Innovation Council, which largely gives out grants to entrepreneurial academic teams and start-ups, will be the biggest winner, with an inflation-adjusted budget boost of around 139%. The fundamental research-focused ERC will see a more modest increase of around 54%. 

In other words, whether this new wave of economic research translates into bigger public science budgets remains to be seen.