Update: China’s PPP R&D Lead Shrinks to 0.3 pp — Business Gap +$100B, EXR Still ~50%

What changed since the NSF State of S&E vintage

Our August measurement-science update closed on a clean NSF/NSB State of U.S. Science & Engineering 2026 headline: worldwide GERD at $3.48 trillion in 2024, China at $1.028 trillion (30%), the United States at $1.009 trillion (29%), and US+China combined at 59%. Publication volume concentrated further so China–US–India alone printed half of world S&E articles. That post answered the first flow question after the 2022 research ledger — did the newest Indicators vintage show the duopoly tightening? Yes.

This Q3 refresh answers the next vintage question the theme brief forces: what changed once OECD Main Science and Technology Indicators (March/April 2026) and the AAAS Global R&D Update (9 June 2026) re-stated the same 2024 year with finer shares, sector splits, and converter sensitivity? Three information events rewrite the frame without overturning the direction of travel. AAAS on OECD prints China at 29.4% of global R&D and the US at 29.1% — an overtake that still holds, but by 0.3 pp instead of the rounded 1 pp gap. OECD’s statistical release warns that GDP PPPs are not R&D PPPs: under alternate converter assumptions China sits at roughly 90–95% of US GERD, and under market exchange rates at about 50%. Sector tables show the private ledger is doing the work — China’s business BERD outspent the US by about $100 billion (constant 2020 PPP) while the US still leads government-sector R&D by about $19 billion.

The dashboard above is built as a measurement delta: share restatement dumbbells, a four-frame China/US ladder, sector-gap bars, OECD growth and EU/US paths, a funding-mix donut, and an AAAS scoreboard for dollars, people, papers, and patents.

The headline table: NSF prior print → OECD/AAAS Q3

Treat the −0.6 / +0.1 pp share moves carefully. They are mostly measurement restatements of the same 2024 year — finer AAAS/OECD percentages versus NSF’s rounded Discovery summary — not a fresh 2025 GERD year. The economically important news is the frame panel and the sector panel: whether China “leads” depends on which converter you trust, and where the dollars sit (business vs government) is more informative than the 0.3 pp headline gap.

The overtake survived — barely — on finer shares

Rounded tables are policy-friendly and journalistically sticky. NSF’s 30% vs 29% print made the China overtake impossible to miss. AAAS’s 29.4% vs 29.1% print keeps the ranking but removes the rhetorical cushion. A 0.3 pp gap is inside the noise band of PPP revisions, coverage differences across non-OECD reporters, and rounding conventions. The duopoly share eases only slightly (59% → 58.5%). Concentration did not unwind; the story got more precise and therefore more fragile.

That is the right reading for the theme’s core question — where is progress funded and published, and is it concentrating? Funding is still concentrating in two economies that together clear roughly three-fifths of the OECD-coverage world. The newest vintage does not restore a clear US lead on current PPP. It does say the lead change is a knife-edge once you stop rounding.

Measurement frames: PPP overtake, EXR near-half, sensitivity in between

OECD’s March 2026 release is unusually explicit about converter risk. Current PPP puts both China and the United States across the $1 trillion mark in 2024 prices, with China slightly ahead. The same release notes that available PPPs are designed for GDP, not R&D, and may be revised in future benchmarking. Illustrative sensitivity in the OECD note — treating four yuan as purchasing the same R&D inputs as one US dollar rather than the latest PPP ratio near 3.5 — places China at about 90–95% of US GERD rather than ~102%. On market exchange rates, China’s R&D is about 50% of the US level in 2024, up from 44% in 2014 and 8% in 2004 — a long catch-up that still leaves the US clearly ahead when bills are paid in dollars.

The frame ladder in the dashboard is therefore not a gotcha chart. It is the honest multi-meter view the prior NSF-only post could not show. Analysts who only quote PPP will keep saying China leads. Analysts who only quote EXR will keep saying the US leads by roughly 2×. Both sentences can be true in the same year. The Q3 contribution is forcing that coexistence into one panel.

EU comparisons reinforce the same lesson. Versus the United States, EU R&D fell from about 70% to 60% on PPP between 2014 and 2024, and from 57% to 43% on exchange rates. The US–China drama is not the only concentration story; the transatlantic gap widened on both converters while East Asia’s intensity leaders (Korea, Taiwan, Japan) keep punching above volume share.

Business dollars flipped the total; government did not

AAAS’s sector cut is the cleanest “where is progress funded” answer in this vintage. Most of China’s rise versus the US is inside the private / business sector, where China outspent the United States by roughly $100 billion in constant 2020 PPP dollars. In the government sector, the United States still outspent China by about $19 billion. Add those signed gaps and you recover the knife-edge total overtake — private strength more than offsets a still-US-led public ledger.

OECD-area funding mix remains business-dominated: about 64% of OECD R&D financing from business (>$1.45 trillion in the AAAS 2023 cut), roughly 23% from government, and the residual from higher education, nonprofits, and other sources. Long-run index growth since 1992 shows the same asymmetry — business financing up about 303%, higher-ed/nonprofit up about 332%, government only about 65%. The newest OECD prose adds a composition twist inside the government slice: budgets are declining in several members while remaining appropriations reorient toward defence. That is a different concentration axis than country shares — mission mix inside the public meter — and it was absent from the August NSF share update.

Growth is uneven: US +3.4%, EU flat, Germany slightly down

Inside the OECD area, 2024 real GERD growth was not a synchronized boom. The United States printed about +3.4%, the European Union about +0.4%, and Germany a slight −0.4% decline. OECD commentary groups Japan, Korea, and Türkiye among economies growing more than 5%. OECD-area R&D intensity remains stuck near 2.7% of GDP from 2020 through 2024 — a plateau that means headline dollar growth is mostly scale and composition, not a broad intensity lift across the club.

The non-OECD ledger is the starkest concentration meter added this quarter: China accounts for 92.3% of non-OECD R&D spending in the AAAS reading of 2024. “The rest of the world outside the OECD” is, operationally, a China story with a thin fringe. Pair that with the prior post’s publication concentration (China–US–India at half of world articles) and the theme’s answer hardens: progress is funded and published in a narrowing set of poles, even when converter debates blur who sits in first place on dollars.

Scoreboard: dollars, people, papers, patents

AAAS’s blunt bottom line — the United States has lost first place across dollars, people, papers, and patents — extends the August pubs-and-GERD story into workforce and IP meters the prior update treated lightly. China continues to widen researcher FTE leadership. PCT patent filings remain China-led since roughly 2012, with ICT patenting especially concentrated (AAAS notes China filing on the order of three times the second-largest ICT filer in 2024). The prior post’s citation-impact caveat still matters: volume and patent counts are not the same as highly cited article share, where the US retained an edge in the State of S&E framing. Q3 does not erase that influence gap; it says the input and output volume scoreboard is now China-first on more columns than the August GERD post alone conveyed.

For fiscal-industrial context on how governments try to bend these curves with place-based and mission budgets, see the fiscal industrial policy Q3 update. For the theme baseline before either update, keep the 2022 research print beside the August NSF vintage.

What would rewrite this update

1. A new GERD year (2025) in NSF Indicators or OECD MSTI that moves China/US shares by more than a rounding restatement.
2. PPP benchmarking revisions that lock the alternate sensitivity (90–95%) into the headline converterflipping the overtake off on official PPP.
3. Sector tables showing the business BERD gap compressing below ~$50B or the US losing the government-sector lead.
4. Non-OECD diversification that pulls China’s share of non-OECD R&D materially below ~90%.
5. GBARD detail quantifying how large the defence reorientation is in US, EU, and East Asian government budgets for 2025–26.

Until those print, the live frame versus the August NSF post is narrow but consequential: the PPP overtake still holds on finer shares (29.4% vs 29.1%), but the gap is only 0.3 pp; EXR and alt-PPP frames reject the overtake; China’s business sector is ~$100B ahead while US government R&D remains ahead by ~$19B; non-OECD R&D is 92.3% China.

Caveats and methodology

1. Same reference year, new converters. Most Q3 deltas restate 2024 GERD, not a 2025 flow year. Do not read −0.6 pp China share as a period decline.
2. NSF rounded shares ≠ AAAS finer shares. Both use OECD-comparable GERD; presentation precision differs. We treat AAAS 29.4 / 29.1 as the newest disclosed fine print.
3. GDP PPPs ≠ R&D PPPs. OECD’s own release flags this; the 90–95% band is sensitivity, labeled estimated in the data module.
4. Constant 2020 PPP sector gaps (business +$100B, government −$19B) follow AAAS’s OECD synthesis; they are not identical to NSF’s current-dollar country totals.
5. Japan/Korea >5% growth is grouped in OECD prose; chart points are estimated midpoints of that “over 5%” statement.
6. Publication and HCA meters from the August post are not fully re-estimated here; the scoreboard cites AAAS directionally for patents and FTEs.
7. Domestic NCSES US R&D totals still must not be mixed into OECD-comparable world shares.

Primary sources: OECD Main Science and Technology Indicators March/April 2026 statistical release; AAAS Global R&D Update 2026 (9 June 2026); NSF/NSB State of U.S. Science and Engineering 2026; prior theme posts measurement-science-update-2026 and measurement-science-research-2026.

The shareable takeaway

Versus the NSF State of S&E print, China’s PPP world share restates to 29.4% against the US at 29.1% — an overtake that survives by only 0.3 pp. Under market exchange rates China is still about half US GERD; under OECD PPP sensitivity it is 90–95%. The private sector explains the knife-edge: China leads business BERD by roughly $100B while the US leads government R&D by about $19B. Non-OECD R&D is 92.3% China. Funding and publication progress are still concentrating — the newest vintage mainly teaches that how you convert decides whether the dollar race looks already lost, nearly tied, or still America’s to call.