Insight, Minerals & Industrial Policy
Zimbabwe will stop lithium concentrate exports in January and has one completed processing plant. Morocco is starting battery production without needing a restriction at all. The difference is not ambition. It is whether the threat is credible and whether the plant is open.
Zimbabwe intends to prohibit the export of lithium concentrates from January 2027, a decision announced in June 2025 by Mines Minister Winston Chitando on the basis that domestic processing capacity would exist by then. Two lithium sulphate plants were named: Bikita Minerals, owned by Sinomine, and Prospect Lithium Zimbabwe, owned by Zhejiang Huayou Cobalt. Zimbabwe holds Africa's largest lithium reserves, and its principal lithium assets are Chinese owned, with Arcadia acquired by Huayou for 422 million dollars in 2021 and Bikita by Sinomine.
The situation now, as reported, is that only one lithium sulphate facility is operational, and it cannot take material from third party mines.
Separately, Gotion Power Morocco's integrated lithium iron phosphate plant at Kenitra is entering production, following an initial 100 million euro African Development Bank package with up to a further 141 million euro to be mobilised, with a first phase of around 10 gigawatt hours of cell and pack capacity and long term plans reaching 100 gigawatt hours alongside cathode manufacturing.
Two countries, the same ambition, opposite instruments. Zimbabwe is using a restriction to force capability into existence. Morocco is building capability next to a customer and has needed no restriction at all.
The comparison is instructive because export restrictions are the most popular industrial policy instrument in Africa at the moment, and their record depends on conditions that are rarely stated.
An export ban is not a tax. It is a threat, and a threat only produces investment if the people who must invest believe it will be carried out.
Kydland and Prescott's account of time inconsistency describes exactly this situation. A policy that is optimal to announce may not be optimal to execute when the moment arrives, and rational actors, knowing this, discount the announcement in advance. The announcement's power comes from the government's ability to bind itself, not from the announcement itself.
Zimbabwe has a revealed preference on this. In 2023 the government required lithium miners to submit plans for local refineries by March 2024, and softened its stance after lithium prices collapsed.
That is the single most important fact in the file, and it is not a criticism of the decision. Relenting was probably correct at the time. But it taught every producer in the country what the government does when enforcing the rule would be expensive, and producers have priced that lesson ever since.
The consequence is a trap with a specific shape. A ban is meant to force investment in processing. Processing plants take two to three years to build and are financed against a decade of operation. So the investment decision is taken years before the ban bites, under uncertainty about whether it will bite, by firms who watched the government blink once already. And the ban is least credible precisely when a plant is least attractive to build, because both depend on the same low price environment.
An announcement made in a strong market and enforced in a weak one requires more institutional commitment than most governments possess, and the ones that have succeeded, Indonesia most prominently, did so by holding the line through several years of complaint and by staging the restriction over roughly six years rather than announcing it once.
Set credibility aside and assume the ban lands on schedule. What happens next depends entirely on a technical detail that almost never appears in the policy discussion.
If the operating lithium sulphate plant is captive to its owner's mine and cannot accept third party concentrate, then a ban does not create a domestic processing market. It creates a monopsony.
Follow it through for a mid-sized Zimbabwean producer with no plant of its own. It may not export concentrate. The only domestic buyer of its concentrate is a plant owned by a competitor, which has no obligation to take the material and every reason to price it at whatever the producer's alternative is worth, which is now nothing.
The value does not move from a foreign refiner to Zimbabwe. It moves from smaller Zimbabwean-operating miners to the two large Chinese owned integrated operators. That is a redistribution among producers inside the country, not a capture of value for the country, and the fiscal position may barely change.
Tolling access is therefore not a technical footnote. It is the difference between an industrial policy and a transfer. A restriction paired with a regulated obligation for processing plants to accept third party material at published, non-discriminatory terms is a fundamentally different instrument from a restriction alone, and it is the version that has a chance of building a supplier base rather than concentrating one.
This is the same governance question we set out about an open battery swapping standard, arriving in mineral processing. Whoever controls the only facility that everyone must use is a regulator whether or not anybody has said so.
Five conditions govern whether an export restriction transfers value to a country or destroys revenue on the way to concentrating it inside one company.
Fail the first three and the policy destroys export revenue. Fail the second alone and it transfers value between domestic producers rather than capturing it for the country.
It is worth being exact about the size of the prize, because the language around beneficiation tends to inflate it.
Lithium sulphate is an intermediate. It is refined onward into battery grade lithium carbonate or hydroxide, which is then converted into cathode material, which goes into a cell. Capturing the sulphate step captures one conversion, and it is a real one with real margin, but it is not entering the battery supply chain in the sense the phrase usually implies. Everything above it, including the qualification relationships that determine who supplies cathode material to whom, remains where it already is.
That does not make the step unworthy. It makes it a step, and describing it accurately matters because the political capital spent on a restriction should be proportionate to the value it captures.
Morocco is not restricting anything. It is building a cell plant next to the largest vehicle market on its border, using an existing automotive manufacturing base, with development bank financing and a Chinese technology partner.
There is a structural logic underneath this that deserves noting. The plant is producing lithium iron phosphate chemistry, which contains no cobalt and no nickel. LFP's cathode inputs are lithium, iron and phosphate, and Morocco holds an extraordinary share of the world's phosphate reserves. A country with phosphate building LFP cathode capability is doing something quite different from a country with lithium ore trying to move one step downstream. It is matching a domestic input to a chemistry that is gaining share globally rather than losing it.
The indicator worth watching is not capacity. It is local integration: how much engineering, component production, workforce capability and supplier development ends up inside Morocco rather than a foreign owned plant simply being located there. That is the capability question, it takes a decade, and it is not visible in a gigawatt hour figure.
Neither approach is obviously superior. Morocco's depends on proximity to a customer that Zimbabwe does not have, and on an existing industrial base that took thirty years to build. Zimbabwe's depends on a resource that Morocco does not have. The point is that the instruments are not interchangeable, and the restriction only works under conditions that have to be assembled first.
For any government considering an export restriction on a mineral, these are the questions to answer before the announcement rather than after.
Is the processing capacity commissioned and operating, or announced?
Construction schedules slip and commissioning takes longer than construction. A restriction dated against a plant that is not yet running is a restriction dated against a forecast.
Can producers without their own plant get their material processed?
On what terms, published where, enforced by whom. If the answer is a commercial negotiation with a competitor, the policy is a transfer.
Do the inputs exist?
Reagents, reliable industrial power, water, effluent handling, and the logistics to move a different product to a different customer. Processing consumes far more of all of these than concentration does.
Is there a buyer for the intermediate?
An intermediate product needs a customer who has qualified the supplier. Concentrate has a liquid market; lithium sulphate has a smaller set of counterparties.
Has the government ever relented before?
If it has, the announcement is discounted, and the discount is largest in exactly the price environment where the policy is most needed. The remedy is a binding mechanism rather than a louder statement: legislation rather than a ministerial announcement, a phased schedule with intermediate milestones, and consequences for the government itself if it moves the date.
Zimbabwe's ambition is legitimate and the frustration behind it is entirely reasonable. A country holding the continent's largest lithium reserves and exporting concentrate to be refined elsewhere is capturing a fraction of the value in its own ground, and the standard advice to be patient has been offered by people who benefit from the current arrangement.
But the restriction is not the policy. The plant is the policy, and the terms on which other people's material can enter that plant are what decide whether the country or two companies capture the difference. A deadline is a useful device for concentrating minds. It is not a substitute for the thing it is a deadline for, and when it arrives before the thing is ready, the government faces a choice between enforcing a rule that destroys revenue and relenting again, which makes the next announcement weaker still.
That is a solvable problem, and it is solved in the eighteen months before the date rather than in the week after it.
The Lab works on this in local manufacturing and across the regions where processing capacity is being built for the first time. The distance between a policy announcement and an operating industrial system is the thing we measure.
If you are assessing a beneficiation programme against a legislated deadline, tell us what you need to know.
See also The Survey Is the First Act of the Mine on how data sequencing is the same instrument acting earlier and more cheaply than an export restriction.
This is an independent insight piece by Transitions Lab. For the Lab's applied work, see Local Manufacturing & Supply Chains. See also The Lock-In Runs Both Ways on qualification as the barrier above lithium sulphate, Who Holds the Pen on the same monopsony question in an open standard, Capability Is the Slow Part on the decade it takes for local integration to appear as capability, and Symbiosis Does Not Arrive on a Site Plan on third-party access to shared infrastructure inside an industrial zone. To discuss a study, see Contact.