Three forms recur, and they differ in law while behaving alike in structure.
A long-term offtake agreement commits a share of future production to a named buyer, often for the life of a mine, sometimes at a formula price set when nobody knew what the decade would hold. A commodity prepayment advances cash against future delivery, so the obligation is settled in tonnes rather than money. A resource-backed loan finances infrastructure against a stream of future commodity revenue, occasionally against the output of the very asset the loan built.
Each is a legitimate instrument with a real economic function. Projects that would otherwise not be financed get financed; a buyer that needs supply security gets it and pays for it. Nothing in this chapter argues otherwise, and nothing in it requires any party to have acted in bad faith.
What the three share is a term structure that outlasts the political cycle of the producing state, and a counterparty relationship that is exclusive in effect whether or not it is exclusive in law.
Brazil exported 416.4 million tonnes of iron ore in 2025 — a record, and the first year above 400 million — for US$28.9 billion. 67% of it went to a single destination. Data
Those figures are real, current, and published. They describe destination concentration, they are computable by anyone with trade data, and they are the kind of number that gets quoted in a ministry briefing as evidence of dependence.
They are also the wrong number for the question that matters.
Destination concentration says where the tonnage went last year. It is a revealed outcome, it can change next year, and a producer facing bad terms can in principle sell elsewhere. Contractual commitment says what the producer is no longer free to sell. It is a constraint, it cannot change until the term runs, and a producer facing bad terms has no elsewhere.
A country can sit at high concentration and low commitment. That is a commercial position: uncomfortable, visible, and reversible within a shipping season. A country can sit at moderate concentration and high commitment. That is a structural position: comfortable-looking, invisible, and fixed for a decade or more.
The two are indistinguishable in trade statistics. They produce the same export tables, the same destination shares, the same ministerial talking points. Only one of them is a trap, and the public record cannot tell you which one you are looking at.
The proposal is narrow on purpose. One country, one commodity, three numbers, published annually.
E is the encumbrance ratio: how much of what the country produces is already promised. C is the concentration of that promise. T is how long until it lifts. No model, no forecast, no opinion, and no price.
The three are not substitutes. E = 0.40 with T = 2 is a rolling commercial book. E = 0.40 with T = 18 is a different country. Reporting either without the other is the same defect this series catalogues elsewhere as a name reaching further than the thing it measures.
Here is the part that makes this political science rather than trade reporting.
Bargaining power is not a stock of resources. It is an option: the credible ability to transact with somebody else. A producer with three plausible buyers and a producer with one plausible buyer can hold identical reserves, identical output and identical costs, and will not obtain identical terms — because the first can leave the table and the second cannot. The option is the whole of the difference.
A long-horizon supply commitment sells that option. Not as a side effect: it is the substance of what the buyer is paying for, and supply security is precisely the name of the thing. The buyer is buying the removal of the seller's alternatives, and the price of the financing is the price of that removal.
The cash is recorded. The delivery obligation is recorded. The option — the future bargaining position extinguished by the term — is recorded by neither party, because no standard requires a seller to value the counterparties it has foreclosed. A country can therefore sell its negotiating position for twenty years and book the proceeds as revenue, with no line anywhere showing what was given up.
Two consequences follow, and both are structural rather than moral.
Foreclosure compounds. Each commitment reduces the volume available to any future counterparty, which reduces the incentive for a new buyer to invest in the relationship, which reduces the number of plausible buyers at the next negotiation. The position does not decay gradually; it ratchets.
Downstream capacity is priced out before it is proposed. A domestic processor competing for committed tonnage is bidding against a contract, not against a market. The familiar pattern of exporting an input and importing the good made from it is usually explained by industrial capability. Commitment is a mechanism that would produce the same pattern with capability held constant, and nobody has separated the two because the data to do so is not published.
The objection is immediate and it is fair: these are commercially sensitive contracts, and parties will say that disclosure damages their negotiating position.
The answer is that the aggregate requires no contract to be published. A country can report that 40% of its output is committed with a weighted remaining term of eleven years, and name no counterparty, no price, no volume by contract, and no party to any agreement. What is disclosed is a property of the national position, not of anyone's deal.
This is not speculative. Extractive-industry reporting established that aggregation of exactly this kind is feasible and that the predicted commercial harm did not materialise; the EITI's own guidance note on resource-backed loans, and the Natural Resource Governance Institute's work on disclosure, are where the argument has already been had. What has not happened is the extension from loans, where disclosure is now a live standard, to offtake, where it is not.
Earlier drafts of this argument ran the chain “iron ore → steel → Brazilian aircraft”. It does not close. Airframes are aluminium alloys, titanium and composites; iron ore is not the input, and a reader in a mining ministry would discard the argument on that line alone. The Brazilian examples that do close are bauxite and alumina, where the input genuinely is the aerospace metal, and niobium, where Brazil's share of world supply is not merely large but close to structural and which goes into precisely the high-strength and superalloy applications advanced manufacturing depends on. That share is not stated here because it has not been verified for this chapter; establishing it is the first item of further work, and if it is what it appears to be it is the strongest single fact this volume has.
The general structure survives the correction unchanged — a country can export the input and import the finished good, and long-horizon commitment is a mechanism that sustains it — but the correction is kept on the page because a chapter that argues for disclosure and quietly fixes its own errors is not making its case.