North Mara: when the dispute is quantum, the valuation file is the whole case

This article concerns a live and contested dispute involving a named operator. Everything attributed to campaign organisations below is an allegation, not a finding; the company's denial is set out alongside it. The Tanzanian and UK proceedings referred to were undetermined at the time of writing, and nothing here should be read as a conclusion about them.

Responding to allegations about relocations for the expansion of the North Mara gold mine, Barrick's position was reported in these terms: there are some 32 cases in the Tanzanian courts concerning the relocations, and they are about how much people agreed to be paid — not about whether they were forced out.[2]

Take that framing entirely at its own word, as a defence, and see what it concedes. It concedes that dozens of households were moved, that compensation was agreed, and that the agreed amount is disputed by enough of them to fill a court list. On that account the case is not about conduct at all. It is about valuation — which means it will be decided on the valuation file.

"It's a quantum dispute" is not the reassuring answer it sounds like. It means your rate schedule is the evidence.

What is alleged, and what is denied

MiningWatch Canada published a report in December 2023 alleging that Indigenous Kuria communities were forced off land for the mine's expansion, with most evictions in December 2022 and the remainder in August and September 2023. It describes the process as "intimidating, coercive and sometimes violent," and states that villagers "were prohibited from using their land to feed and support themselves long before they received any compensation."[1][2] The report attributes consequences including landlessness, loss of food security, children withdrawn from school and dispersal of families among neighbours for close to a year. It follows correspondence the organisation had sent the company in January 2023.[3]

Barrick denies that the evictions were forced, its chief executive rejecting that characterisation directly.[2] Separate UK proceedings concerning the mine, on matters other than resettlement, have been brought on behalf of Tanzanian claimants.[4]

I am not in a position to resolve that disagreement and do not try to. What is useful to a practitioner is the structural point underneath it, which holds whichever account is right.

The government valuer problem

The valuation underlying the relocations is reported to have been prepared by the Ministry of Lands, Housing and Human Settlements Development, covering some 652 acres, in June 2021.[1]

That is the ordinary Tanzanian route: statutory compensation runs through the government valuer, and market value with no depreciation plus the section 179 allowances is the domestic measure.[5] For a project financed or certified against lender standards, that route creates a structural gap that has nothing to do with anyone's good faith:

• The government valuer's mandate is the statutory measure, not PS-5 replacement cost. Where they diverge — depreciation, transaction costs, the ability actually to acquire an equivalent asset — the statutory figure can be entirely correct and still short of the lender test.[6][7]
• The valuer is not the operator's, so the operator frequently holds no independent working papers — no comparables, no field measurements, no rate derivations. When quantum is litigated, the company's evidence is a report it commissioned from a third party and cannot itself explain.
• A single valuation exercise covering hundreds of acres and many households creates correlated risk: if the method is challenged successfully once, it is challenged successfully thirty-two times.

The mitigation is unglamorous and cheap relative to the exposure: run a parallel replacement-cost assessment against the statutory one, record both, pay the higher, and keep the derivation of each. Then a quantum dispute is answered with working papers rather than with a defence of someone else's report.

Expansion is where mining resettlement actually goes wrong

A mine is not a road. Its footprint is not fixed at approval — it grows with the pit, the waste rock dumps, the tailings facility and the buffer zones, over a life measured in decades. The resettlement consequence is that a community can be acquired from repeatedly, in separate exercises, each internally defensible and each using whatever rates and standards applied in its own year.

Two failure modes follow, and both are records problems:

• Cumulative loss goes uncounted. A household that gave up grazing land in one phase, a field in the next and its homestead in a third has been compensated three times at market value and may still be destitute — because each exercise measured a parcel and none measured the household's trajectory. Only a register that persists across phases, keyed to the household rather than to the acquisition, will show it.
• Standards drift. The rates, the eligibility rules and the applicable lender policy in a 2011 exercise are not those of a 2021 exercise. Unless each phase records the standard it was run against, later claimants comparing outcomes across phases will see arbitrariness — and the project will have no way of demonstrating otherwise.

The sequencing allegation is the one to design against

Of everything alleged, the most consequential for a project's own file is the claim that people were barred from using their land before compensation reached them.[1] Whether or not it happened here, it is the allegation most easily made and most easily rebutted — and the rebuttal is a date.

Karuma is the cautionary authority: Uganda's Court of Appeal awarded general damages for the constitutional breach of taking possession before payment, independently of the valuation dispute, which the claimants lost. A project that records, per household, the date access was restricted and the date funds became available has a complete answer. A project that cannot produce those two dates is defending an allegation with recollection.

What the register had to carry

• Two valuations per asset — statutory and replacement-cost — with the derivation of each and the top-up where they diverge.
• Access-restriction date against funds-available date, per household, evidenced.
• A household identity that survives across acquisition phases, so cumulative loss over the mine life is visible as one record rather than three unrelated ones.
• The standard each phase was run against, named and dated, so differences between phases can be explained rather than merely observed.
• Consent evidence separated from payment evidence. A signed acceptance proves a document was executed. Whether the amount was understood, and whether refusal was a real option, are different propositions needing different records.

What this case does not establish

Nothing above is a finding that any eviction was unlawful, that any compensation was inadequate, or that the operator acted improperly. The allegations are those of MiningWatch Canada and are attributed as such; the company denies the characterisation of the relocations as forced, and that denial is cited from the same source that carries the allegations.[2] The acreage, dates and the June 2021 Ministry valuation are as reported in that campaign material and have not been verified against a project register or against the valuation report itself, to neither of which this article had access. The Tanzanian relocation cases and the UK proceedings were undetermined at the time of writing. This is not legal advice on Tanzanian land law or on any claim arising from this project.