Page 2 of 10
The two monopolies
China supplies most of the world's rare earths with, by its own count, only about a quarter of world reserves: in its first white paper on the industry, in June 2012, Beijing said it met over 90% of global demand with 23% of world reserves.
Geological conditions alone do not explain the monopoly. Decades of policy choices in Beijing do.
The story is usually told from 1979, but it starts earlier: Baotou Steel was built with Soviet help in 1954, and by 1963 Baotou had its own rare-earth research institute. China's first monopoly, though, was a monopoly of raw materials. From 1979, Deng's reforms opened the sector to foreign money but never to foreign ownership: by 1990 rare earths were a protected strategic mineral, and export-tax rebates from 1985 pushed volume abroad. The West did not resist. By 1999–2000 the United States got more than 90% of its rare earths from China, and most mines outside China had closed, Mountain Pass in 2002 among them. Cheap supply let Western countries keep the dirtiest stages of the chain far from home.
The second monopoly is different. From 1999 Beijing began restricting its own output with export quotas, production caps and environmental rules. After losing a WTO case in 2014 it moved control inside the industry: two state champions, quotas only for them, and since 2023 a ban on exporting magnet-making technology.
Today China mines about 60% of the world's magnet rare earths but refines about 91% of them, importing feedstock, much of it from Myanmar, to fill the difference. It also makes about 94% of the world's sintered permanent magnets, up from around half two decades ago.
The puzzle
One country holds most of the chain, and the chain is political
In its 2012 white paper, China's State Council said the country "supplies over 90 percent of the global market rare earth needs" with 23 percent of world reserves. Klinger (2017) calls production sites "geologically contingent, rather than geologically determined". Once a commodity is labelled strategic, governments act on it. China has used quotas, export duties, industry consolidation, technology bans and licensing since 1999. Washington, Brussels and Tokyo now reply with their own sovereignty policies.
If dominance was built by policy, can policy move it? Or do policies mostly put a label on what trade is already doing?
The chain Data · all BACI flows
China left the cheap end of the chain and took the expensive end
Rare earths are a chain, not one product. We follow four stages through six trade codes, with ferrite magnets as a placebo that rare-earth policy should not touch. Pick a stage to see how China's share of world exports has moved.
China's exports ÷ world trade at each stage, CEPII BACI HS92 V202601, 1995–2024, all bilateral flows. Descriptive only: this is not yet a test of H1, and it is not the dependence measure of H3, which uses each importer's own sourcing. 8505.11 also contains non-rare-earth metal magnets.
Guess first
In 2024, what share of the world's permanent-magnet exports (HS 8505.11, by value) came from China?
Exports are not production: China makes about 94% of the world's permanent magnets but exports a smaller share, because it uses many of them at home.
Policy Draft event list · coding provisional
Fifty-seven policy moves, 1999–2026
Our test depends on telling two kinds of policy apart. Binding instruments (quotas, licences, bans) change what can legally leave a port. Declaratory ones (white papers, five-year plans) say what a government wants. We expect only the first kind to break trade series. Click any marker for its source.
Source: group list of policy events (60 events, 41 verified against primary text; coding provisional). The binding / declaratory / hard-case coding is provisional. Shaded bands are the five periods of the text analysis. Two pre-1999 events and one undated event are not plotted.