Changing lithium ownership may not secure supply, model suggests

Changing who owns a lithium mine does not necessarily change where its ore ends up. A new study predicts that divesting Chinese stakes in overseas lithium producers would leave global lithium output almost unchanged, whereas banning ore shipments to China would sharply reduce supply at existing capacities. The distinction is between changing commercial incentives and cutting access to processing facilities that other countries cannot immediately replace.

The findings come from Yale economist Kei Irizawa’s Cross-Border Ownership and Input Access: Evidence from the Lithium Supply Chain, posted as an arXiv preprint on November 5, 2026. These are simulations from an estimated economic model, not measured effects of implemented divestitures or export bans.

Ownership matters, but nationality is not enough

The paper’s data show the geographical mismatch: in 2024, China accounted for 72% of refinery output but 21% of mine output. Allocating production by equity stakes raises Chinese owners’ share of world mining output to 38%.

Irizawa assembles a global operator-site panel for 2015–2024, primarily from Benchmark Mineral Intelligence and S&P Capital IQ Pro, supplemented by company filings and public records. It covers production, capacity, costs and ownership, alongside market prices and regional demand. Quantities are standardized to lithium carbonate equivalent, a measure of lithium content.

A separate descriptive analysis examines 1,485 potential links between 33 producing hard-rock mines and 45 refineries with complete supplier lists. Sharing an owner is associated with a roughly 60-percentage-point higher probability of a reported supply relationship, after accounting for mine and refinery characteristics, host-country pairs and distance. Chinese ownership on both sides adds no positive association once those factors are included. This supports distinguishing corporate affiliations from common nationality; it does not establish that ownership causes sourcing decisions.

How the model works

The model separates hard-rock mines, which sell concentrate for refining, from integrated brine producers, which make lithium without purchasing traded ore. Mines choose spot-market commitments first. Refiners then choose sourcing and output, while brine producers choose output in the same final lithium market.

Equity stakes determine how much producers value profits elsewhere in their owners’ portfolios. A refiner affiliated with a mine shares in the mining margin, lowering its effective input cost even though affiliated and spot purchases carry the same modeled price. Conversely, a mine may limit spot sales that would strengthen competing refiners. Shared ownership among lithium producers can also discourage expansion that would reduce rivals’ profits.

Demand and costs are estimated jointly using generalized method of moments. The demand equation controls for GDP, nickel and oil prices, regional differences and a time trend. Lagged global hard-rock mining capacity instruments for lithium prices; rival capacity and ownership-linked mining-cost shocks help identify cost relationships. The estimated demand elasticity is −0.511, with a standard error of 0.172: locally, a 1% price increase implies about a 0.51% consumption decline, holding demand shifters fixed.

Two restrictions, very different outcomes

The policy exercises use 2024 conditions, holding capacity and entry fixed while allowing sourcing, production and prices to adjust. Divestiture transfers Chinese stakes in overseas hard-rock mines and integrated brine operations to separate local owners without existing lithium interests, leaving trade open.

Against a re-solved observed-ownership benchmark, spot ore purchases rise 20.93%, while ownership-linked purchases fall 47.84%. More spot trading does not mean cheaper inputs: the ore price rises 0.76%. Hard-rock lithium output falls 1.90%, but brine output rises 3.17%, leaving total lithium output just 0.06% higher. Chinese refiners can still buy internationally.

An alternative exercise hypothetically pools Chinese owners’ refining interests, including foreign brine holdings. Divestiture then raises lithium output 2.22% and reduces spot purchases 2.72%, each relative to its own pooled benchmark. Stronger brine expansion reduces demand for traded ore. This is a sensitivity exercise about competition, not evidence of actual coordination.

The export-ban scenario retains ownership but blocks overseas ore deliveries to China through both spot and affiliated routes. Global lithium output falls 25.36%, and the world lithium price rises 77.19%. Ore becomes cheaper outside China and dearer inside it. Hard-rock refineries outside China reach capacity; overall, production outside China replaces only 41.0% of China’s lost output.

What the estimates cannot settle

The predictions depend on proportional influence from equity, equity-based purchasing limits, profit-maximizing conduct and a homogeneous global lithium market. Contractual pricing differences are simplified. The supplier data report relationships, not shipment volumes or transaction prices. Instrument validity also requires capacity and cost shifters to be unrelated to the relevant unobserved shocks after controls.

Model fit is informative but imperfect: predicted mine output has an R-squared of 0.851 across 139 mine-years, yet aggregate mining output is overpredicted by 17.25%. The reported policy tables provide point estimates, not uncertainty intervals.

Finally, fixed capacity excludes investment responses that could change longer-run outcomes. These stylized interventions do not replicate individual national policies or quantify the security benefits governments might seek. The narrower lesson is that changing ownership does not reserve ore for domestic users, and redirecting ore cannot substitute for the capacity needed to process it.