What happened to mineral prices after export controls?

14 of the 29 rows in this table, from 12 of the 23 controls, have a price twelve months after the month before the control. In 9 of them the price was higher, and the same 9 rose by more than the median of all priced commodities over the same months. The median change was +11%, against a market-wide median of +4%. 5 of the 14 start in 2025, so a single year's market move can drive them; without them, 9 rows remain, the price was higher in 4, and the median change was −2% against +3%. With so few cases, a few large market moves decide the pattern, and nothing here shows that a control caused a price change.

Each row is a dated entry in the export-controls register that names a mineral with a benchmark price series. Some entries ease or suspend earlier controls; those rows are marked, and the status shows measures that have ended. The base is the average price in the last full month before the announcement month, so the change at 3, 6 or 12 months is measured from that base month. Beside each change are two reference points: where it falls among every change of the same length in that price's own history (from 2000, or from when the series begins), as a percentile where 90 means only a tenth of those changes were larger, and the median change of all priced commodities over the same months, which is the market-wide move. Entries with status scheduled or unconfirmed are left out, and so are three that never took effect: a Senate bill that was dropped before enactment; controls suspended before their start date; a proposed duty that was withdrawn and never in force.

By horizon

Every row

Why this cannot show cause

• Few, overlapping rows. 23 controls produce 29 rows because several name more than one mineral. Several come from one country in the same months, or follow each other (a suspension, its extension and then quotas), so the rows are not independent.
• Some rows ease a control. A partial lifting or a suspension of earlier controls is marked on its row; it points the other way from a ban, and the table does not treat it differently in the arithmetic.
• Prices have many drivers. Demand, other supply, currencies and stocks move prices at the same time. The market-wide median is a rough control for common moves, not for the mineral's own, and it includes food and energy as well as minerals.
• Dates are approximate and the base month can hold the news. The register date is the announcement or legal act; markets may have reacted earlier on reports, a measure may start later or be eased, and for some entries the announcement falls inside the base month.
• The benchmark is often not the controlled good. Each mineral uses one benchmark series, usually for the refined metal, while many controls cover ore, concentrate, scrap, processing technology or a licensing rule. The row says what the entry controls.
• Some series barely move. A benchmark that is administered or reported at a near-fixed price shows almost no change for months (less than 0.1% a month); rows where that happened inside the window are marked, and a zero change there is not a market result.
• Different histories. Percentiles use the series' own history, which begins in 2012 for lithium and manganese and in 2000 or earlier for the rest, and they include changes that end after the control as well as before it.
• No price series, no row. Controls on minerals without a free price series on this site do not appear, and some series end a month before the latest, which removes the windows that would end in the latest month.

Export-controls register (this site) and World Bank Pink Sheet and IMF benchmark prices. Prices to September 2026 for the series that run that far. Generated 2026-10-22. Data: event-prices.json.