Charted: US + China Fund 57% of Global R&D — Science Concentrates

Science is often narrated as a borderless commons — open journals, shared methods, collaborative labs. The funding and publication ledgers tell a tighter story. In 2022, gross domestic expenditure on R&D (GERD) across the OECD-coverage world totaled about $3.1 trillion in purchasing-power-parity dollars. The United States performed 30% of that stack and China 27%. Together, the two largest systems fund 57% of measured global research. The top eight individual performers — adding Japan, Germany, South Korea, the UK, France, and Taiwan — account for 82%.

That is the core answer to this post's question: where is progress funded and published, and is it concentrating? Funding is concentrating into a US–China duopoly sitting on a thin high-income tier. Publication volume has concentrated even faster toward Asia: China alone produced 27% of peer-reviewed science and engineering articles in 2022, versus 14% for the United States, after overtaking US article counts around 2016. Citation impact, however, still favors established producers — volume leadership is not the same as influence leadership.

The interactive dashboard above is the map: a GERD share streamgraph, a publication-rank bump chart, 2022 dollar volumes, an impact-versus-volume scatter, and R&D intensity bars. Pair the global research ledger with our OECD DAC aid drop — another concentration story about who pays for public goods — and with who owns AI compute, where capital concentration in a different stack (chips and data centers) is even sharper.

What GERD actually measures

GERD is the Frascati-manual total of R&D performed inside an economy in a year — business, government, higher education, and private nonprofits — converted for cross-country comparison with PPP. NSF's Science & Engineering Indicators, drawing on OECD Main Science and Technology Indicators, is the framing used here. A few measurement rules matter before any chart:

1. PPP dollars are not market-exchange dollars. China's GERD looks closer to the US in PPP than it does at market rates; both comparisons are used in the literature, and they answer different questions.
2. The "world" total is OECD-coverage, not every UN member. NSF's 2022 global figure (~$3.1T) covers OECD members plus a set of non-member reporters including China and Taiwan. Missing economies bias the denominator slightly.
3. US GERD in OECD tables is adjusted for international comparability (capital treatment, depreciation, some foreign-funding adjustments) and can differ from domestic NCSES totals printed elsewhere in the same reports.
4. Mid-path share years between disclosed anchors (2000, 2010, 2022) are estimated for shape. Treat the streamgraph as a concentration narrative, not a basis-point time series.

WIPO's end-of-year R&D blog puts 2023 spending near $3 trillion in current dollars, with the US near $784B and China near $723B. Those are useful current-dollar cross-checks; they are not mixed into the PPP share charts here.

The concentration table

Read the table as one sentence: research funding is a duopoly-plus-tier story; publication volume is an Asia-shift story; citation impact is still a quality-weighted story.

The funding stream: from US plurality to US–China parity

In 2000 the United States performed roughly two-fifths of measured global R&D. China's share was a mid-single-digit footnote. By 2010 the US share had already fallen to about 31% as China's climb and Korea's intensity pushed the Asian share up. By 2022 the US is still first at 30%, but China is a near peer at 27%. The EU-27 as a bloc is 18% — larger than Japan (~$201B, roughly 6–7%) but no longer a rival to either superpower on dollars alone.

Absolute volumes make the duopoly visceral. NSF/OECD 2022 GERD prints put the United States at about $923 billion PPP and China at about $812 billion. Japan ($201B), Germany ($175B), and South Korea ($139B) are the next tier — each less than a quarter of China's stack. The UK, France, and Taiwan round out the top eight. Toggle the dashboard's US–China duopoly view: the two leaders' combined share rises from the mid-40s in 2000 toward the high-50s by 2022, while the "rest of world" ribbon compresses.

Concentration here is not "everyone stopped researching." Global GERD grew. What concentrated is share: a larger fraction of a larger pie sits in two systems. That matters for standards-setting, talent magnets, dual-use technology races, and the bargaining power of research universities and corporate labs inside those systems.

For a parallel look at how growth and trade redistribute economic weight — the demand side that funds much business R&D — see our macro growth, trade & prices research.

Publication volume: China leads; the pack reshuffles

Peer-reviewed article counts are an early output meter of the research system. NSF's publications framing (Scopus-based fractional counts) shows China at 27% of world S&E articles in 2022 and the United States at 14%. India is third at about 6%. Germany, the UK, and Japan each sit near 3%. Six countries combined clear half of global output.

The bump chart is the cleanest visual of the flip. The United States led volume until roughly 2016, after which China took first place and never looked back. Japan fell from a top-two producer in the early 2000s toward sixth by 2020 as its article counts stagnated while China's rose roughly tenfold from 2003–2022 and India's about eightfold. US output grew only about a third over the same window — respectable in absolute papers, catastrophic as a share.

Field mix matters. In 2022, health sciences dominated US, UK, German, and Japanese portfolios; engineering dominated China's; computer and information sciences dominated India's. Counting "papers" without field weights is like counting ships without distinguishing freighters from destroyers — useful for scale, incomplete for capability.

Impact is not volume: the HCA caveat

Citation impact still separates the story. NSF's highly cited article (HCA) index sets the world baseline at 1.0. Around 2020 the United Kingdom printed the highest large-producer HCA near 2.2; the United States and Germany were near 1.7. China's index has been climbing toward and around baseline as volume exploded — progress, but not yet US-style citation leverage per paper. India's fast volume growth still sits below baseline on impact in the same framing.

The dashboard's impact-versus-volume scatter puts that tension on one plane: China far right on share, mid-pack on HCA; the UK and US higher on HCA with smaller (UK) or mid (US) volume shares. Caveat: HCA is field- and database-sensitive; open-access shifts and language coverage change who looks "highly cited." Treat the scatter as a directional warning against equating article counts with scientific influence — not as a precise league table of national genius.

Intensity: small systems punch above their weight

Dollar shares reward large economies. GERD as a percent of GDP rewards commitment. In 2022 Israel led at about 6.0% of GDP and South Korea at 5.2%. Taiwan printed near 4.0%. The United States sat near 3.6%, Japan 3.4%, Germany 3.1%, China 2.6%, and the EU-27 about 2.1%.

China's intensity is still below the US even while its PPP dollar total rivals the US. That is the arithmetic of a still-poorer-per-capita but enormous economy running a sustained industrial-science strategy. Korea and Israel show the other extreme: mid-sized economies that treat R&D intensity as a national survival parameter. Intensity leaders are often the same places that show up in advanced manufacturing, semiconductors, and defense-tech ecosystems — see also our IRENA renewable capacity record for how capital-intensive technology buildouts concentrate geographically once learning curves and industrial policy lock in.

Who funds inside the big systems

Across the top performers, business is the largest performer and usually the largest funder of R&D. NSF notes that for five of the top eight, businesses perform more than three-fourths of R&D. Government funding shares in the high teens to low twenties of GERD appear in Korea, the UK, the US, and China. Higher education's performing share is double-digit mainly in China, Germany, and France among the largest systems with OECD detail.

Industry focus diverges. US business R&D leans hard into information and communication services and pharmaceuticals; Germany and Japan put large slices into motor vehicles; the UK and France put more into scientific R&D services; aerospace shares are notable in France and the US. Those sectoral fingerprints explain why "national R&D" debates so often collapse into chip fabs, auto platforms, or drug pipelines — the ledger is already specialized.

Is it concentrating — and what that means

Yes on funding share and publication volume; partially on impact; no on the idea that only two countries do science. The rest of the world still performs hundreds of billions of dollars of research and millions of papers. What changed is the center of gravity: a late-20th-century US-and-allies plurality has become a US–China bipolar funding core, with Europe and Japan as large but relatively shrinking share holders, Korea and Taiwan as intensity specialists, and India as the fastest-rising publication volume story after China.

Policy implications follow without a sermon:

• Talent and immigration policy in the US and Europe is a research-capacity lever, not only a labor-market lever.
• Research security and collaboration rules now operate inside a bipolar funding map; decoupling cuts collaboration bandwidth where the papers and labs actually are.
• Quality metrics (citations, reproducibility, standards) matter more as volume floods; raw article counts are a poor scoreboard for national strategy.
• Allied intensity (Korea, Israel, Taiwan, parts of Europe) is a force multiplier for open scientific ecosystems even when absolute GERD is smaller.

Caveats and what we are not claiming

We are not claiming Chinese science is "better" or US science is "done." We are claiming the budget and bibliography have concentrated. We are not equating GERD with innovation outcomes (startups, patents, productization). We are not treating PPP and market-exchange rankings as interchangeable. Mid-path streamgraph and bump years are estimated between disclosed anchors. Publication databases undercount some languages and overcount some conference-heavy fields. HCA indices move with database vintages.

The honest headline remains quantitative and narrow: in 2022, two countries funded 57% of measured global R&D, eight countries funded 82%, and six countries wrote more than half the papers — with China leading volume since 2016 while citation impact still favors older scientific powers.

Sources and further reading

Primary framing: NSF NCSES / National Science Board Science & Engineering Indicators (Discovery R&D chapter; State of US S&E 2024 publications section; Publications Output reports). Cross-checks: OECD MSTI GERD tables; WIPO Global Innovation Index R&D blog for 2023 current-dollar levels. Method notes are in the dashboard source line and in the measurement section above.