Charted: Subsidy Jurisdictions Nearly Doubled as Industrial Policy Went Global

Industrial policy never left. What changed is who uses it, how often, and which instruments do the work. The IMF’s collaboration with the Global Trade Alert — the New Industrial Policy Observatory (NIPO) and its Historical extension (H-NIPO) — turns that anecdote into a countable ledger: 34,248 distinct interventions from 2009 through 2023, of which 85% were trade-distorting. China, the European Union, and the United States alone account for roughly 53% of the historical stock.

The interactive dashboard above is built for the question this theme asks: how do subsidies, tariffs, and industrial policy reshape economies? The short answer in the data is that subsidies remain the workhorse, tariff-style trade tools are rising inside advanced-economy toolkits, and since 2020 the motives have shifted toward supply-chain resilience, national security, and geopolitics — a shift Global Trade Alert’s late-2025 briefing describes as a turn toward more coercive, externally oriented industrial policy.

Unlike our US industrial subsidies vs tariffs post — which tracks dollar outlays and customs revenue on the US federal ledger — this piece tracks policy activity worldwide: how many jurisdictions intervene, with which instruments, and under which stated motives. Dollar packages still matter; they appear in the fiscal-war-chest panel. Counts matter more for tit-for-tat dynamics.

The extensive margin: more countries subsidize

The headline is not that every capital suddenly discovered industrial policy. It is that subsidy use nearly doubled at the extensive margin: from a little over one-third of monitored jurisdictions in 2009 to nearly three-fifths in 2023. Localization policies doubled from 6% to 12%. Trade barriers as a jurisdiction share look flat across endpoints — until you look at 2020, when pandemic emergency measures pushed import/export tool use into the majority of both advanced and emerging economies.

Toggle the coverage metric in the dashboard. Subsidies tell a structural story. “Any distorting IP” tells a crisis-and-persistence story: the share hit 75% in 2020 and again in 2022, then eased to 63% in 2023 — still above the 2009 baseline. The post-pandemic world did not revert to the 2010s.

Inside the toolkit: AE–EMDE convergence on subsidies

Instrument shares within industrial-policy measures tell a different story from jurisdiction coverage. Advanced economies entered the sample already subsidy-heavy: 84% of AE industrial-policy measures in 2009 were subsidies, falling to 75% by 2023 as trade measures rose from 3% to 8% and localization edged up. Emerging and developing economies moved the other way: subsidies rose from 56% to 71% of their IP measures, while trade measures fell from 27% to 18%.

That is convergence toward a subsidy-led global norm, not a return to 1970s-style tariff walls as the primary tool — even as tariffs re-enter Western politics. The GTA’s 2025 “Security First” briefing warns that early-2025 data show Western interventions tilting back toward coercive trade tools relative to domestic financial support. The H-NIPO 2009–2023 ledger captures the long climb of subsidies; the 2025 briefing captures the next turn of the screw.

For US readers who only watch CHIPS and IRA press releases, the AE trade-share rise from 3% to 8% looks small. It is not small in political economy terms: it is the thin edge of export controls, local-content rules, and trade-defence actions sitting inside the same industrial-policy category as grants.

The 2023 wave: 2,580 measures, 71% distorting

The original NIPO snapshot for calendar 2023 recorded 2,580 interventions: 98 plans and strategies (3.8%), 1,451 policies and regulations (56%), and 1,031 firm-specific awards (40%). Roughly 71% were trade-distorting. China, the EU, and the US again dominated — about 48% of that year’s measures.

Two implications follow. First, industrial policy is mostly implementation, not communiqués: firm awards and regulations outnumber strategy documents roughly twenty-five to one. Second, a subset of 882 import-restrictive measures with identifiable trade coverage touched at least 22% of global trade. That is not a niche steel case; it is a measurable wedge through the trading system.

Pair this count-based view with our China fiscal revenue breakdown when you care about how another major industrial-policy actor actually books revenue on the fiscal ledger. NIPO counts selective interventions; budget breakdowns show the cash room that makes subsidy races feasible.

Motives: from climate and competitiveness to security

Before 2020, competitiveness and climate dominated stated motives in the H-NIPO narrative. After 2020, supply-chain resilience, national security, and geopolitics move to the foreground. Inside the national-security / geopolitics motive class, the instrument mix itself mutated:

• Export barriers rose from 7% to 22% of those measures.
• Import barriers fell from 28% to 11%.
• Localization / procurement fell from 30% to 9%.
• FDI measures fell from 24% to 11%.
• Other / unconventional instruments jumped from 5% to 37%.

Read that carefully. Security-motivated industrial policy did not simply “add tariffs.” It reweighted toward export controls and unconventional tools — technology denial, outbound investment screens, entity lists — while classic localisation and import-barrier shares within that motive declined. GTA’s 2025 briefing adds that more than half of recent US industrial-policy measures explicitly cite national security or geopolitics, with security-of-supply motives pushing Western interventions toward a security-centred regime that is harder to reverse once framed that way.

Fiscal packages: the war chests behind the counts

Counts do not show fiscal scale. The dashboard’s package panel puts statutory and mobilisation headlines side by side:

• US CHIPS Act semiconductor appropriations: $52.7B (including $39B for manufacturing incentives).
• CHIPS advanced manufacturing ITC: often scored near $24B, with upside scenarios far higher.
• European Chips Act: roughly €43B (~$47B) mobilised public-and-private investment targetnot a single EU cheque.
• EU IPCEIs for microelectronics, batteries, and hydrogen: on the order of €37B+ (~$40B) approved state aid.
• IRA clean-energy tax expenditures: original headline near $370B, with independent uptake estimates often much larger.
• China’s Big Fund III, Japan, and Korea semiconductor packages: tens of billions each on reported envelopes.

These figures are not comparable outlays. Appropriations, tax-credit scores, state-aid approvals, and equity-fund raises answer different budget questions. They are comparable as political commitments: every major manufacturing power is putting nine- and ten-figure claims on semiconductor and clean-tech geography. Our commercial aircraft final-assembly map shows what decades of subsidy competition look like in a mature dual-use industry; chips and batteries are earlier innings of the same game.

Who is exposed — and what would change the story

Exposed: export-oriented manufacturers in jurisdictions that cannot match subsidy intensity or that sit downstream of new export controls; WTO dispute settlement, which was not designed for simultaneous security-framed industrial policies across dozens of capitals; climate coalitions that assumed green industrial policy would stay cooperative rather than become a subsidy race with local-content tripwires; EMDE governments that converge on subsidy tools without AE-scale fiscal space.

Relative winners under current rules of the game: firms that can collect overlapping grants, credits, and procurement preferences across CHIPS/IRA/IPCEI-style regimes; capitals that combine fiscal capacity with export-control leverage; sectors already inside strategic product lists (semiconductors, critical minerals, dual-use, low-carbon tech).

What would change the story: a sustained drop in new NIPO interventions back toward pre-2019 rates; a negotiated subsidy code that bites on security-framed measures; or a fiscal crunch that forces AE governments to choose between industrial credits and other mandatory spending. None of those appear in the 2023–2025 ledgers.

Caveats and methodology

• Counts ≠ dollars. One firm-specific award and one multi-year strategy both count as interventions; fiscal impact differs by orders of magnitude.
• Trade coverage is a lower bound. The 22% of global trade figure applies only to the 882 import measures with identifiable coveragenot the full distorting set.
• Intermediate coverage years are interpolated between disclosed endpoints and peak years; confidence flags in the data file mark disclosed vs estimated points.
• Package USD figures mix metrics (appropriations, mobilisation targets, state-aid totals, tax-credit scores) and FX conversions; treat them as order-of-magnitude war chests.
• EU is often coded as a single actor in NIPO summaries even though member-state aid drives much of the fiscal reality.
• 2025 instrument-mix claims from GTA’s Security First briefing are directional early-year evidence, not a full-year census matching the 2023 NIPO table.
• H-NIPO uses LLM-assisted motive tagging for 2009–2022; misclassification risk is real even when aggregates are robust.

The shareable takeaway

From 2009 to 2023, the share of jurisdictions using trade-distorting subsidies rose from 36% to 59%, while 34,248 industrial-policy interventions accumulated — 85% of them distorting, and more than half from China, the EU, and the US. The 2023 wave alone logged 2,580 measures, and security-framed policy increasingly reaches for export barriers and unconventional tools. Subsidies still dominate the toolkit; tariffs and controls are how the same agenda becomes coercive.

Related reading: US industrial subsidies vs tariffs (30 years) and China fiscal revenue breakdown.