A puzzle stumps statisticians: How much does China actually spend on R&D?

To many, statistics can seem a dull business. But in the past few months, an odd dispute between Beijing and Paris over R&D statistics has perked up the profession.

At issue is measuring how much China spends on R&D – a question that touches on how much Western politicians should worry about being overtaken by Chinese technology.

In May, the Paris-based Organisation for Economic Cooperation and Development, which maintains a central database of science and technology statistics for much of the world, quietly posted notice online that it is “suppressing" publication of annual Chinese R&D statistics dating back to 2019. The reason: “anomalies” in data reported by Beijing to the organisation.

The anomalies involve China’s report to the OECD in February about its R&D budgets and workforce.  Among several “inconsistencies” the OECD says it found: numbers that suggest China’s non-manufacturing industry boosted its annual spending on R&D by a whopping 99% during 2020, even as its manufacturing R&D slipped by 1.4% - an improbable combination in the same year. And for 2019 China reported a huge 42% jump in the population of higher-education researchers despite a budget increase of only 23% to pay for them.

OECD officials say they haven’t yet received a detailed explanation from Beijing, and decline to comment beyond the lengthy note they already published on their statistics website in May. (For convenience, Science|Business reprints here the full statement, which can be hard to find on the OECD site).

But a Chinese official – speaking, he says, in his personal capacity as an expert in science and technology statistics rather than as a spokesman for his government – told Science|Business that the problem may be partly due to some mistakes in some of the numbers China reported.

“Discussions are underway” with the OECD, and the “China side is checking the data in the OECD information note,” says Changlin Gao, minister counselor for science and technology at the Mission of China to the EU. “But I don’t know how long it will take” to resolve the issue.

The politics of numbers

But why the fuss, amplified on social media in the small global community that follows these kind of statistics? In part, it’s because the OECD rarely halts publication of so many statistics from so important a country as China. And in part, it’s because China’s rising R&D investment – however big it is – is anxiously watched by tech policy experts in Washington, Brussels and other Western capitals.

By any measure, China’s R&D power has soared in this century. The country’s total R&D spending jumped 11-fold from the year 2000, to hit $464 billion, or 2.14% of gross domestic product, in 2018, the most recent year for which the OECD is still publishing Chinese numbers. It blew past EU spending of $377.8 billion or 2.07% of GDP, and came close to US spending of $586 billion, or 3.01% of GDP, in that year. Since then, the Biden Administration has cranked up US R&D spending faster still, to $707.8 billion in 2021 or 3.46% of GDP. (The dollar figures are adjusted by the OECD to account for national differences in costs – so-called purchasing-power parity numbers.)

But statisticians have long been aware that different countries count science and technology indicators differently – and it can get very political. For instance, just last year Britain’s statistics office announced it was revising its numbers for small-business R&D expenditures. Though the office denied any political motive, the result was a convenient justification for the government to cut its tax subsidies for R&D spending as part of its overall efforts to restrain post-COVID spending.

In the case of China, it’s even harder to reconcile numbers because the Chinese economy is structured so differently than that in the Western democracies.

"China doesn’t count it (R&D) the way that OECD countries count it,” says Caroline Wagner, a professor at Ohio State University. For instance, in China, big tech companies have more government control than in the West, so it isn’t always clear how much of their R&D spending is government or private sector. Likewise, Western analysts have trouble categorising R&D spending by the Chinese Academy of Sciences.

“All of these things makes it really hard to compare” with the West, says Wagner.

Apples and oranges

Of course, the OECD tries to make sense of all this. Among its functions is compiling statistics from its 38 member-states, plus “partner” countries like China. For its science and technology statistics, it has detailed number-counting guidelines in its Frascati Manual, named after the Rome suburb at which experts from around the world first met to try comparing statistical apples and oranges in 1963. The manual, says Gao, has been translated into Chinese and the government follows it in its reports to Paris.

But in this case, Gao says, he understands there were “some mistakes” in China’s latest Yearbook of Science and Technology Statistics, and they are being corrected.

For instance, the OECD’s note says China reported a 10% rise for 2020 in overall business enterprise R&D, to 1.87 trillion renminbi ($261 billion), including a slight drop to RMB 1.47 trillion for manufacturing industry. But that, the OECD says, implies a 99% rise in spending for the rest of the business sector. In fact, says Gao, the Chinese yearbook numbers were wrong: he estimates that manufacturing R&D actually rose rather than fell, to between RMB 1.6 trillion and RMB 1.65 trillion – and that fact, rather than any unusual jump in other economic sectors, is what contributed to the overall 10% rise in the country’s business-enterprise R&D.

Another issue highlighted by the OECD is how many researchers China has in its higher education sector. It says China’s numbers for R&D personnel suggest a 42% jump in 2019, but a budget increase of only 23% to pay for them. That makes sense, Gao says, because research costs in Chinese higher education are generally lower than in Europe – so the budget increase doesn’t have to match the personnel increase. And in another area of the statistics, Gao also calls “unprofessional” the OECD’s critique of some inconsistencies between three different measures of public spending in China. American numbers, Gao says, sometimes show similar patterns.