A food group buys palm oil, cocoa and maize through a handful of large traders. It knows those traders well, holds contracts with all of them, and can produce an invoice for every tonne. It cannot name a single farm. The emissions it needs to disclose happen on land it has never seen, three relationships away.
This is the industry where the disclosure requirement and the available evidence are furthest apart, and where the usual advice to improve supplier engagement does not survive contact with the supply chain.
Key points
- Land use change is a direct emissions source in this industry, alongside soil management and fertiliser, not only a Scope 3 item.
- Agricultural Products is not one of the nine industries the ISSB prioritised for review. Processed Foods is, and the two sit next to each other in the same sector.
- Proposed amendments would remove the Ingredient Sourcing topic entirely and replace the supply chain topics with environmental supply chain management provisions.
- Supplier-level evidence is the binding constraint. The tier where the land sits is usually the tier with no contractual relationship to the reporting entity.
Land use change is not a Scope 3 footnote
The most common mistake in this sector is filing land use change under Scope 3 and moving on. The Agricultural Products standard treats it differently. Direct emissions in this industry arise from soil management practices such as fertiliser application, from land use change including deforestation and land conversion, from processing, and from transport.
Read that list again. Land conversion sits alongside fertiliser as a direct source.
The distinction that matters is who farms the land. Where the entity farms it, conversion emissions are its own, in Scope 1. Where a supplier farms it, they arrive through purchased goods in Scope 3 category 1. Same physical event, different scope, decided by the ownership of the operation rather than the nature of the emission.
That is why a plantation company and a food manufacturer buying from that plantation report the same hectare in entirely different places, and why neither can check the other's number.
Where this industry sits in the ISSB review
Paragraph 32 of IFRS S2 requires an entity to refer to and consider the applicability of the industry-based metrics. Applying them is not a condition of stating compliance.
Be precise about the review status, because the two industries are easy to confuse. In July 2025 the ISSB proposed a comprehensive review of nine prioritised industries: all eight in Extractives & Minerals Processing, plus Processed Foods. Agricultural Products is not among them.
It is affected all the same. Proposed amendments would remove the Ingredient Sourcing disclosure topic and the environmental and social impacts of ingredient supply chain topic, along with their metrics, and replace them with environmental supply chain management provisions. Those proposals are still at exposure draft stage.
So a preparer here is in an awkward position. The topics most relevant to your hardest disclosure are the ones proposed for replacement, and the replacement is not final. Build the underlying supplier record rather than the current metric format, because the record survives a redefinition and a metric mapping does not.
Why supplier data thins out exactly where it matters
Traceability in agriculture degrades predictably by tier.
Your direct suppliers are contracted, invoiced and named. You can ask them for data and they have a commercial reason to answer.
Below them sit mills, crushers and aggregators. These buy from many farms and blend the output. A mill can often tell you its throughput and sometimes its catchment, but the physical product you bought is a mixture by the time it reaches you.
Below that are smallholders. You have no contract with them, no purchase order, and frequently no list of who they are. In several commodities they supply the majority of volume.
The awkward arithmetic is that emissions intensity is usually highest at the tier where your evidence is weakest. Land conversion happens at the farm, not the mill. So the entity is asked to disclose a figure whose largest component sits furthest from anything it can document.
This is not solved by a supplier questionnaire. A questionnaire reaches tier one, and tier one is not where the land is.
Why spend-based factors fail here
When physical data is missing, the usual fallback is a spend-based factor: multiply what you paid by an emissions factor per unit of currency. It is a reasonable approximation for office supplies or professional services.
It behaves badly for a traded commodity. Agricultural prices move on weather, currency and speculation, none of which change the emissions of growing the crop. A drought that doubles the price of a tonne of cocoa doubles your reported emissions for buying the same tonne. Your footprint rises in a year when your physical purchasing did not change at all.
Use physical volume where you have it, which in this industry you usually do, because tonnes are what the contracts are written in. Keep spend-based factors for the categories that genuinely have no physical unit, and disclose where you used them.
What the entity can actually hold
Given that, the useful question is not how to obtain farm-level data. It is what record you can defensibly keep instead.
Three things are obtainable. Volume by commodity and origin, from your own purchase records. The methodology and factors you applied to convert that volume into emissions, with the source and vintage of each factor. And the basis on which you allocated a blended supply to an origin, which is a judgement and should be recorded as one.
Those three, kept properly, let you reproduce the figure. Farm-level primary data, chased and not obtained, does not.
Malaysian reporters have a nearer-term version of this problem, since palm oil supply chains run through exactly this tier structure and the Bursa reporting timetable is already running. The structure of the problem is the same everywhere the commodity is grown by many small producers.
What a preparer has to be able to show
An assurance provider will sample a commodity and follow it back. Expect four questions.
What volume did you buy of this commodity in the period, and does it tie to purchase ledgers. What origin did you assign it, and on what basis, given the supply is blended. Which emission factor did you use, from which source, of which vintage. And where the figure rests on an assumption about supplier practice, who approved that assumption.
The first is usually solid, because it is a finance number. The second is where it breaks. Origin allocation is often done once, in a spreadsheet, by someone reasoning sensibly from trade flows, and the reasoning is never written down. A year later the number can be recalculated but the judgement cannot be explained.
The fourth barely exists in most first cycles. Assumptions about supplier practice get made implicitly by choosing a factor, and nobody records that a choice was made at all. That is the same weakness that shows up whenever a figure has to be reproduced later, concentrated here because so much of the number is assumption rather than measurement.
In this industry the evidence file is mostly a record of judgements. Treat it that way from the start, and name the person behind each one.
Common questions
Is land use change Scope 1 or Scope 3?
It depends on who operates the land. Where the reporting entity farms it, conversion emissions are direct and sit in Scope 1 alongside soil management and fertiliser. Where a supplier farms it, they reach the entity through purchased goods and services in Scope 3 category 1.
Is Agricultural Products part of the ISSB review of priority industries?
No. The nine prioritised industries are the eight in Extractives & Minerals Processing plus Processed Foods. Agricultural Products is a neighbouring industry in the same sector, and separate proposed amendments would change its supply chain topics, but it is not one of the nine.
What happens to the Ingredient Sourcing topic?
Proposals would remove it, together with the environmental and social impacts of ingredient supply chain topic and their metrics, replacing them with environmental supply chain management provisions. This is at exposure draft stage and is not in force.
How should smallholder supply be evidenced?
Through what you can hold rather than what you cannot obtain: purchase volumes by commodity and origin, the factors applied with their source and vintage, and a written record of how blended supply was allocated to origin and who approved that allocation.
Where do you stand against IFRS S2?
The readiness diagnostic scores governance, evidence and controls. The IFRS S2 reference sets out the standard paragraph by paragraph, and the deadline checker gives the first reporting period for your market.