Dispute Resolution in Commodity Trading: Contracts, Evidence, and the Governed Record

Executive summary

Every commodity trading organisation, however well run, will eventually find itself in a dispute. A counterparty reads a contract differently, a delivery arrives off-specification, an invoice is challenged, a force majeure is declared, a price is contested, or a position is unwound in a way one side considers a breach. Disputes are not a sign of failure; they are an inherent feature of a business built on contracts, physical delivery, volatile prices, and long chains of counterparties. What separates organisations that handle disputes well from those that suffer through them is not the absence of disputes but the quality of their preparation, their evidence, and their process.

This guide sets out how commodity trading disputes actually arise, the mechanisms through which they are resolved, from direct negotiation through mediation and arbitration to litigation, and the practical steps a trading organisation can take to resolve them faster, more cheaply, and more often in its favour. Its central argument is simple: the single greatest determinant of how a dispute resolves is the quality of the record. A trading operation that can produce a complete, timestamped, tamper-evident history of a trade, from the moment it was agreed through every amendment, valuation, nomination, and settlement, holds an enormous advantage over one that must reconstruct events from emails, spreadsheets, and memory.

That is why dispute resolution, which looks like a legal topic, is in large part a data and systems topic. The governed, auditable record that a modern ETRM platform maintains is not only an operational asset; it is the best insurance policy a trading business can hold against the disputes it cannot avoid. This guide is written for the people who live with that reality: heads of trading and operations, legal and contracts teams, risk managers, and the technologists who build the systems of record they all depend on.

Why disputes are inherent to commodity trading

To manage disputes well, it helps to understand why they are so common in this industry specifically. Several features of commodity trading combine to make disagreement almost inevitable over a large enough book and a long enough time.

The first is that commodities are physical. Unlike a purely financial instrument that settles in cash against an agreed reference, a physical commodity has to be produced, stored, transported, and delivered, and every one of those steps introduces the possibility that what arrives is not exactly what was promised. Quality can drift outside specification. Quantity can be measured differently at load and discharge. Delivery can be late, or to the wrong location, or interrupted by an event neither party controls. Each physical step is a potential source of dispute that simply does not exist in a cash-settled world.

The second is that prices are volatile. When a contract is struck, both parties are content with its terms. By the time it performs, the market may have moved enormously, and the party now on the wrong side of that move has a powerful incentive to find a reason the contract should not bind, or should bind differently. Volatility does not create disputes on its own, but it sharpens them: a disagreement about a delivery window matters far more when prices have doubled since the trade was struck, because the money at stake in the resolution is far larger.

The third is that contracts are complex and often incompletely specified. A commodity contract has to address quality specifications, measurement, delivery point and window, title and risk transfer, payment terms, credit support, force majeure, governing law, and the mechanism for resolving disputes, and it frequently incorporates by reference a set of standard terms, an industry master agreement, or a set of general terms and conditions. The interaction between the specific deal terms and the incorporated standard terms is a rich source of genuine ambiguity, and ambiguity is where disputes live.

The fourth is the length and opacity of the chain. A cargo may be bought and sold many times in transit; a molecule of gas may pass through several counterparties between producer and burner. When something goes wrong far down a chain, the dispute can propagate back up it, and each link may have documented the same underlying events slightly differently. The more hands a trade passes through, the more versions of the truth exist, and the harder it is to establish what actually happened.

Taken together, these features mean that disputes are not an aberration to be eliminated but a cost of doing business to be managed. The organisations that thrive are not those that somehow avoid disputes but those that have built the contracts, the records, and the processes to resolve them efficiently.

The common categories of dispute

Although every dispute is particular, most fall into a handful of recognisable categories, and knowing them helps an organisation prepare for the disputes it is most likely to face.

Quality and specification disputes arise when delivered product does not meet the contractual specification, or when the parties disagree about whether it does. A crude cargo with a higher sulphur content than specified, a gas stream outside its calorific range, a metal shipment that fails assay, each can trigger a claim for rejection, a price adjustment, or damages. These disputes turn on measurement and sampling: who measured, when, by what method, and whether the procedure followed the contract.

Quantity and measurement disputes arise because a physical commodity is measured more than once, at load and at discharge, by different parties using different instruments, and the numbers rarely agree exactly. Contracts allow for tolerances and specify which measurement governs, but disputes still arise about losses in transit, the calibration of meters, and the correct application of temperature and pressure corrections.

Delivery and performance disputes concern whether delivery happened as the contract required: on time, to the right place, in the right manner. A missed delivery window, a failure to nominate, a vessel that arrives outside its laycan, a pipeline nomination that is cut, each can be a breach, and each can trigger claims for the cost of covering the shortfall in a moved market.

Payment and settlement disputes arise when an invoice is challenged, a payment is withheld, or the parties disagree about the amount owed after adjustments for quality, quantity, demurrage, or other charges. These are often the most document-intensive disputes, because resolving them requires reconstructing the entire commercial history of the trade.

Force majeure and change-of-circumstance disputes arise when one party claims that an event beyond its control, a natural disaster, a government action, an infrastructure failure, excuses its non-performance. These disputes turn on the precise wording of the force majeure clause, whether the event genuinely falls within it, whether notice was given correctly and on time, and whether the party took reasonable steps to mitigate.

Valuation and margin disputes arise, particularly in financial and cleared trading, when the parties disagree about the mark-to-market value of a position, and therefore about the collateral that should be posted. In a fast-moving market a margin call can be disputed simply because the two sides value the same position differently, which is itself often a consequence of using different curves or models.

Credit and default disputes arise when one party becomes insolvent or fails to perform, and the surviving party must close out positions, apply netting, and claim under any credit support. The enforceability of netting and the correct calculation of a close-out amount are frequently contested, especially when large sums turn on the valuation methodology used at the moment of default.

The escalation ladder: how disputes are resolved

Commodity disputes are resolved through a well-established ladder of mechanisms, escalating in formality, cost, and finality. A well-drafted contract specifies this ladder in advance, and a well-run organisation understands where on it a given dispute sits and where it is heading.

Direct negotiation is the first and by far the most common resolution. Most disputes are settled commercially between the parties, often between the original traders or their managers, without any formal process. Negotiation is fast, cheap, private, and preserves the commercial relationship, which is why it resolves the great majority of disagreements. Its success depends heavily on both parties being able to see the facts clearly, which is precisely where a good record earns its keep: a negotiation conducted from a complete, shared understanding of what happened is far more likely to settle than one conducted from competing reconstructions.

Mediation is the next step: a neutral third party helps the disputants reach their own settlement, without imposing one. Mediation is voluntary and non-binding, and its value is in breaking a deadlock that direct negotiation could not, by giving each side a structured, confidential forum and a skilled facilitator. Because the mediator does not decide, mediation preserves the parties’ control over the outcome, and it is markedly cheaper and faster than the formal mechanisms that follow.

Expert determination is a specialised mechanism common in commodity disputes that turn on a technical question, a quality specification, a measurement method, a valuation. The parties agree to refer the narrow technical question to an independent expert whose decision is binding. It is faster and cheaper than arbitration for the right kind of dispute, because it puts the question to someone with the domain knowledge to answer it directly rather than to a tribunal that must be educated.

Arbitration is the dominant formal mechanism in international commodity trading. The parties agree, usually in the contract, to submit disputes to a private tribunal whose decision, the award, is binding and enforceable. Arbitration is favoured over litigation in this industry for several reasons: it is private, it allows the parties to choose arbitrators with commodity expertise, it produces awards that are enforceable across borders under the New York Convention, and it is generally faster than court litigation. Many trade associations run their own arbitration schemes with rules tailored to their commodity, and contracts routinely incorporate them.

Litigation in the national courts is the final rung. It is used when there is no arbitration agreement, when the dispute involves parties or issues outside an arbitration clause, or when a party seeks a remedy only a court can give, such as an urgent injunction to freeze assets or preserve evidence. Litigation is public, can be slower and more expensive, and produces a judgment enforceable through the courts. For all its drawbacks it remains essential for certain disputes, and the mere possibility of it shapes the negotiations that usually settle before it is reached.

The important point about the ladder is that most disputes settle low on it. The formal mechanisms are expensive and slow, and both parties usually prefer to avoid them. But the shape of the settlement reached low on the ladder is powerfully influenced by what each side believes would happen high on it, and that belief, in turn, is shaped by the strength of each side’s evidence.

Evidence is everything: the role of the record

Whatever mechanism resolves a dispute, the outcome turns on evidence. Negotiators settle in the shadow of what a tribunal would find; mediators work with the facts each side can establish; arbitrators and judges decide on the record put before them. In every case, the party that can produce a clear, complete, credible account of what happened holds the advantage. This is the single most important practical insight in dispute management, and it is where trading technology meets legal outcome.

Consider what a commodity dispute actually requires the parties to establish. What were the exact terms of the trade as agreed, including every amendment? When was each obligation performed, and by whom? What was the measured quality and quantity at each point, and by what method? What notices were given, when, and in what form? What was the position worth at the moment of default, on what curve, using what model? What did each party know, and when did they know it? Every one of these questions is answered from records, and the quality of the answer depends entirely on the quality of the records.

Now consider the two ways an organisation can hold those records. In the first, the trade lives in a front-office system, its amendments in emails, its confirmations in a shared drive, its valuations in spreadsheets, its nominations in an operations tool, and its settlement in a back-office system, each with its own copy, its own timestamps, and its own gaps. Reconstructing the history of the trade for a dispute means assembling this scattered, inconsistent evidence months or years after the fact, hoping the pieces agree and that nothing critical was lost. This is slow, expensive, and dangerous, because every inconsistency between the records is a weakness the other side will exploit.

In the second, the trade lives once, on a governed model, and every event in its life, capture, amendment, valuation, nomination, settlement, is recorded against that single record with a timestamp and an actor, in a history that cannot be silently altered. Reconstructing the trade for a dispute is a query, not an archaeology project, and the account it produces is internally consistent because it comes from one source. This is the difference a governed, auditable record makes, and in a dispute it is often decisive.

What a governed, auditable trade record provides

The properties that make a trade record valuable in a dispute are specific, and worth naming, because they are exactly the properties a modern platform is designed to provide and a fragmented legacy landscape is not.

Completeness. A record is only as good as what it contains. A governed model captures the full lifecycle of the trade, not just its economic terms but every subsequent event, so that nothing material to a dispute lives outside the record in an email or a spreadsheet. When the whole history is in one place, there are no gaps for the other side to fill with their own version.

Immutability and versioning. A record that can be silently edited is worthless as evidence, because the other side can always argue it was changed after the fact. A governed model versions every change: the original terms are preserved, each amendment is a new version with its own timestamp and actor, and nothing is overwritten. The history is append-only, so the record shows not only what the trade is now but what it was at every past moment and exactly when it changed.

Bitemporality. The most powerful evidential property is the ability to answer questions as of a past moment, both as the world was and as the system knew it. What did we believe the position was worth on the day of the alleged default, using the curve that was live then? A bitemporal record answers this exactly, reproducing the past state rather than approximating it, which is precisely what a tribunal needs when the dispute turns on a valuation at a specific historical instant.

Lineage. When a number is challenged, the ability to trace it to its inputs is decisive. A governed model with lineage can show that a disputed valuation was produced from these positions, this curve, this model version, this configuration, each of which can itself be traced to its source. Lineage turns an assertion (the position was worth X) into a demonstration (here is exactly how X was computed, from auditable inputs), and a demonstration is far harder to dispute than an assertion.

An audit trail of actors and time. Disputes frequently turn not only on what happened but on who did it and when. A complete audit trail records every consequential action, a booking, an amendment, a notice, an approval, against the person or system that performed it and the precise time, in a tamper-evident log. This is the evidence that establishes the sequence of events, and the sequence is often the heart of the dispute.

These properties are not a coincidence; they are the natural output of the same governed, event-driven, bitemporal architecture that makes a trading platform operationally superior. The record that lets a desk reconcile by construction rather than by nightly batch is the same record that wins disputes, because both depend on representing each fact once, completely, and immutably.

Preventing disputes before they start

The best dispute is the one that never happens, and much of dispute management is really dispute prevention. Several practices, most of them supported directly by good systems, reduce both the frequency and the severity of disputes.

Precise, consistent contracts. Many disputes are born in ambiguous or inconsistent contract terms. Capturing contractual terms in a governed, structured form, rather than only in free-text documents, reduces the ambiguity by forcing precision and by making the terms consistent across every deal that incorporates them. When the delivery window, the quality specification, and the governing terms are structured data rather than prose, they are harder to misread and easier to enforce.

Validation at capture. A large class of disputes traces back to an error made at the moment of trade capture: a wrong counterparty, an unvalidated instrument, a delivery point that does not match the contract. Validating trades against governed reference data at the point of entry catches these errors before they become disputes, because the system refuses to book a trade that does not tie out to the master data.

Timely, accurate confirmations. Confirming a trade promptly and accurately with the counterparty is one of the most effective dispute-prevention measures, because it surfaces any disagreement about the terms while it is trivial to fix, rather than after performance when it is expensive. A system that generates and matches confirmations automatically closes the window in which a mismatched understanding can harden into a dispute.

Clear notices and deadlines. Many disputes turn on whether a notice, of a claim, a force majeure, a default, was given correctly and on time. Tracking the notices and deadlines associated with each contract, and alerting when a deadline approaches, prevents the loss of rights that comes from missing a contractual clock, and creates a record that the notice was given properly.

Consistent valuation. Valuation and margin disputes are far less likely when both parties value from transparent, well-understood methods. Internally, valuing every position from one governed model on consistent curves removes the internal disagreements that make external disputes harder to resolve, because the organisation speaks with one voice about what a position is worth.

Running a dispute well when it happens

When prevention fails and a dispute arises, a disciplined process improves the outcome. The organisations that handle disputes best treat them as a managed process rather than a scramble.

Preserve the record immediately. The moment a dispute is foreseeable, the relevant records should be preserved, and a governed system with immutable history makes this automatic rather than a frantic effort to stop people editing spreadsheets. The completeness and integrity of the record at the moment the dispute crystallises is what the whole resolution will rest on.

Establish the facts from one source. Before taking a position, assemble the complete history of the trade from the governed record: the terms and every amendment, the timeline of performance, the valuations and their inputs, the notices and their timing. A clear, internally consistent account of the facts is the foundation of both a strong negotiating position and a strong case if the dispute escalates.

Assess the position honestly. With the facts established, assess the strength of the position candidly, including its weaknesses. The purpose of the record is not to manufacture a case but to understand the real one, so that the organisation can decide whether to settle, and on what terms, or to fight, from a position of knowledge rather than hope.

Choose the right mechanism. Match the dispute to the resolution mechanism: negotiate where a commercial settlement is achievable, use expert determination for a narrow technical question, invoke arbitration or litigation only where the stakes and the strength of the case justify their cost. The contract usually dictates the available mechanisms, but within them there is judgement about how far to escalate.

Preserve the relationship where it matters. Many counterparties are repeat partners, and the value of the ongoing relationship often exceeds the value at stake in a single dispute. A resolution that wins the dispute but destroys the relationship may be a net loss, which is why negotiation and mediation, which preserve the relationship, are so often preferable to the adversarial mechanisms that damage it.

How Gravitas supports dispute resolution

Gravitas is not a legal system, and it does not resolve disputes. What it does is provide the governed, auditable record on which good dispute resolution depends, as a natural consequence of its architecture rather than as a bolted-on feature.

Because every trade lives once on a single governed model, the complete history of a trade, its terms, every amendment, every valuation, every nomination, every settlement, is held in one place rather than scattered across systems and inboxes. Because that history is versioned and bitemporal, any past state of the trade or the book can be reproduced exactly, including what a position was worth on a specific historical day using the curve that was live then. Because every consequential action is recorded in an immutable audit trail against its actor and timestamp, the sequence of events, so often the heart of a dispute, is established rather than reconstructed.

The same properties that prevent disputes are built in too. Trades are validated against governed reference data at capture, catching the errors that become disputes. Confirmations surface term mismatches early. Valuation runs from one model on consistent curves, so the organisation values its positions with one voice. And the analytical layer makes the whole history queryable, so assembling the facts for a dispute is a query rather than an investigation.

In short, the record that makes Gravitas operationally strong, complete, immutable, bitemporal, lineage-tracked, is the same record that makes an organisation strong in a dispute. That is not a separate feature; it is the same governed architecture viewed through a legal lens. For the deeper architectural case, see our cloud-native ETRM whitepaper, and for the operational foundation, how trade capture creates the governed record.

A worked example: the contested close-out

To see how these pieces fit together, follow a common and high-stakes dispute from beginning to end: a contested close-out following a counterparty default. The example is illustrative, but every step maps to a real evidential question.

A counterparty fails to post margin and is declared in default. The surviving party must close out the portfolio of trades with that counterparty, calculate a single net close-out amount, and claim it, applying any collateral held. The dispute, when it comes, is almost always about the number: the defaulting party’s administrators contend that the close-out amount was calculated too aggressively, on unfavourable curves, at a moment chosen to maximise the claim.

Resolving this dispute favourably requires the surviving party to establish several things precisely. First, the exact population of trades in the netting set at the moment of default, which requires a complete, versioned record of every trade and amendment up to that instant. Second, the enforceability of netting across that population, which requires the governed contractual terms linking the trades to an enforceable master agreement. Third, the value of each trade at the close-out moment, which requires the ability to reproduce the market as it was at that instant, the curves, the volatilities, the correlations, exactly as they stood. Fourth, the methodology used, which must be shown to be the contractual one, applied consistently, and traceable from result to inputs.

An organisation running a fragmented landscape struggles with every one of these. The trade population must be reconstructed from several systems that may not agree. The valuation as of the default moment must be recomputed from market data that may have been overwritten. The methodology must be evidenced from spreadsheets. Each gap is a point the administrators will press, and each inconsistency weakens the claim.

An organisation on a governed, bitemporal model answers each question with a query. The netting set at the default instant is the versioned trade population as of that timestamp. The valuation is reproduced exactly because the market snapshot as of that instant is preserved. The methodology is evidenced by lineage from the close-out amount back through the positions, curves, and model version that produced it. The claim is not an assertion to be defended but a demonstration to be presented, and demonstrations settle disputes.

This is the concrete value of the governed record in the dispute that matters most, the one where the largest sums turn on a valuation at a single historical instant. The architecture that lets a desk manage risk in real time is the same architecture that lets it defend a close-out years later, because both rest on the ability to know exactly what was true, and what was believed, at any past moment.

Contracts, standard terms, and governing law

A word is warranted on the contractual foundations, because the mechanism that resolves a dispute is almost always determined in advance by the contract, long before the dispute arises. Understanding these foundations is part of managing disputes well.

Most commodity contracts incorporate standard terms by reference, an industry master agreement or a set of general terms and conditions published by a trade association, an exchange, or one of the parties. These standard terms address the recurring issues, quality, measurement, delivery, title, payment, force majeure, so that the individual deal need only specify what is particular to it. The interaction between the incorporated standard terms and the specific deal terms is both a source of efficiency and, when they conflict or leave gaps, a source of dispute. Capturing which standard terms apply to each trade, as governed data rather than as an assumption, removes a common ambiguity.

The contract also specifies the governing law and the dispute-resolution mechanism, the arbitration scheme and seat, or the courts, that will apply if the parties cannot settle. These choices matter enormously to how a dispute unfolds: different laws treat force majeure, damages, and the duty to mitigate differently, and different arbitration schemes have different rules and different pools of arbitrators. A trading organisation should know, for every material contract, which law governs and which mechanism applies, and a governed record that captures these terms makes that knowledge available rather than buried in a document.

Finally, the contract sets the notice requirements and deadlines that so often decide disputes: the time within which a claim must be brought, a force majeure declared, a defect notified. Rights are lost by missing these clocks, and tracking them against each contract, with alerts as deadlines approach, is one of the highest-value dispute-prevention measures an organisation can take.

Conclusion

Disputes are an unavoidable part of commodity trading, produced by the physical nature of the goods, the volatility of prices, the complexity of contracts, and the length of the counterparty chain. They cannot be eliminated, but they can be managed, and the quality of that management is a real source of competitive advantage: the organisation that resolves disputes faster, more cheaply, and more often in its favour keeps more of what it earns and preserves the relationships that let it keep trading.

The through-line of this guide is that dispute management is, to a surprising degree, a data and systems discipline. The mechanisms of resolution, negotiation, mediation, expert determination, arbitration, litigation, all turn on evidence, and the quality of the evidence is set long before the dispute arises, by the quality of the record the organisation keeps. A governed, immutable, bitemporal, lineage-tracked record is the best insurance a trading business can hold against the disputes it cannot avoid, and it is the same record that makes the business better run day to day. Investing in that record is investing in both operational excellence and dispute resilience at once, because they are, at bottom, the same investment.