EU-Mercosur: Deforestation
Deforestation in Mercosur is its own subject. It has its own causes, its own scale, its own politics, and its own contested literature. Most of what is happening to the forest is not about the EU-Mercosur trade agreement — it predates the deal, and it would continue without it. That is a different piece, and it deserves its own series, which it will get later this year.
This piece is narrower: it is about what we can and cannot say about the relationship between the EU-Mercosur agreement and forest loss. The conversation around the deal is full of confident claims. Ratification will accelerate deforestation. Ratification will reduce deforestation, through the leverage of EU standards. The deal includes meaningful protections. The protections in the deal are paper-only.
A note on vocabulary. The conversation around the deal speaks of deforestation, because that is the term the deal addresses. The ecological question is broader: the conversion of natural habitats, which includes the savannah, grassland, and wetland ecosystems that fall outside the legal frame. This piece uses the narrower vocabulary where the law requires it, and the broader vocabulary where the science does.
What I want to do here is walk through five questions and explain why each of them is harder to answer than the conversation suggests. This piece is an attempt to map the limits of what we are entitled to claim. Mapping uncertainty is a less common register than asserting conclusions. It is also, on this topic, the only honest one. The confident claims that have driven this conversation have, in most cases, outrun the evidence the speakers had access to. Slowing down to ask what we would actually need to know is the work this conversation has been skipping.
1. Compared to what?
The simplest claim about the deal is that it will increase agricultural exports from Mercosur to the EU, and those exports are linked to deforestation. The deal will therefore increase forest loss. The simplest counter-claim is that the deal includes environmental commitments, that the EU Deforestation Regulation will apply to its imports anyway. The deal will therefore reduce or hold steady the deforestation footprint of EU consumption.
Both claims require a comparison. The deal will increase or reduce deforestation compared to what? The honest answer is that the counterfactual either is not specified in these claims, or when it is specified, the specification is contested.
The Commission's counterfactual rests on the claim that the EU is the buyer with the highest environmental standards, and that EU non-ratification would simply redirect supply to less demanding markets.
The premise is partly true. China imports more Brazilian soy than the EU does. But there is another contributing factor: Chinese soybean demand from Brazil surged after 2018, when the United States imposed tariffs on Chinese goods and Beijing retaliated by shifting agricultural purchases away from the American Midwest.
The implicit counterfactual in most NGO analyses is a scenario where the deal does not enter force and Mercosur exports remain at their current levels.
These are different counterfactuals. They produce different estimates of the deal's marginal effect, and choosing between them requires assumptions about global trade flows that are themselves contested.
The range of confident estimates illustrates the difficulty. The French government's own modelling estimates that facilitated market access for Mercosur beef could increase deforestation by at least 5% per year over six years. The European Commission, in its own communications, asserts that the deal "will not generate further deforestation in Mercosur." Two estimates from European institutions: a substantial annual increase, and no effect at all. They do not disagree about Brazilian land use. They disagree about what the deal does to it — which is the question the counterfactual is supposed to answer.
This is the simplest example of why the counterfactual question is harder than the conversation admits. The deal does not enter a static world. It enters a world in which the largest variables driving Mercosur agricultural exports are being reshaped by actors whose decisions in the next decade will affect the deforestation outcome more than the deal's tariff schedules will. The confidence with which the deal's deforestation effect is asserted is harder to defend once the counterfactual is examined.
2. What does the EUDR actually do?
The most common response to deforestation concerns about the deal is that the EU Deforestation Regulation applies anyway. Adopted in 2023, the EUDR requires that seven commodities — cattle, soy, palm oil, wood, cocoa, coffee, and rubber, plus their derived products — entering the EU market come from land that was not deforested after 31 December 2020. The regulation places the due diligence obligation on EU operators and traders, and it applies regardless of whether the exporting country has a trade agreement with the EU. The Commission's defence of the Mercosur deal rests, in significant part, on this regulation existing.
The defence assumes the regulation works as designed. There are problems with this assumption.
Postponements. The EUDR's original implementation date was 30 December 2024. It was postponed to the end of 2025 under pressure from trading partners and EU member states. It was postponed again, to the end of 2026, under further pressure. In each postponement, the regulation was also adjusted — most notably, traceability requirements and shared supply-chain responsibility have been weakened relative to the original text. The current rule is not the rule the Commission's defence of the deal originally referenced.
Definition of a "forest." The EUDR defines "forest" narrowly: lands with an area greater than 0.5 hectares, trees taller than five metres, and forest cover greater than 10 per cent. The Brazilian Amazon meets this definition. The Cerrado — the savannah biome that has become Brazil's primary soybean expansion frontier — largely does not. It is questionable whether the regulation that is supposed to backstop the Mercosur deal's impact on expanding agricultural land into natural habitats applies to the ecosystem where the pressure is actually rising.
Enforcement. The EUDR's due diligence framework places the burden on EU importers to verify that their supply chains are deforestation-free, but the verification depends on traceability systems in producer countries whose maturity varies considerably. The deal will enter provisional application on 1 May 2026. The EUDR's enforcement architecture will, by the regulation's current schedule, become operational eight months later. Whether it functions as intended is a question that can only be answered after the implementation.
Confident claims about what the EUDR will achieve are claims about a regulation whose final shape, scope, and enforcement remain undetermined.
3. How does a tariff change reach the forest?
The deal reduces tariffs on a defined list of Mercosur agricultural exports to the EU. Lower tariffs raise the effective price Mercosur producers receive for exports to the EU. Higher prices make production more profitable. More profitable production attracts investment, expands output, and increases the area of land brought into agricultural use.
Each step of that chain is a modelling decision. Each step has been challenged in the published literature.
Price-transmission assumption: how much of a tariff reduction passes through to the farmgate price the producer actually receives, as opposed to being absorbed by intermediaries — traders, processors, exporters — operating between the farm and the EU border. The literature suggests the pass-through rate for Brazilian commodities ranges from around 30% to over 80% depending on the commodity, the region, and the time horizon. The same tariff reduction can therefore produce very different farmgate effects depending on which estimate the model uses.
Supply-elasticity assumption: how much do producers actually expand output in response to a given price increase. Brazilian and Argentine agriculture have substantial spare capacity in some regions and binding constraints in others. The same price signal can produce expansion onto frontier land, intensification on existing land, or no response at all, depending on where the farm sits in the regional supply curve.
Land-use elasticity: of the expansion that does occur, what share comes from converting forest, what share from converting other ecosystems, and what share from intensifying existing pasture or cropland. This is the assumption with the widest range of estimates in the literature, because it depends on local factors — land tenure, road access, enforcement of the Forest Code, credit availability — that vary across municipalities and that the models tend to handle through aggregated regional averages.
The three sources of uncertainty multiply rather than add. A modelling exercise that varies all three within their plausible ranges produces a forest-loss estimate that can differ by a large factor. This is not because the science is incompetent. It is because the chain from tariff to forest is genuinely long, and each link is genuinely uncertain.
Confident claims about how the deal's tariff changes will affect Brazilian land use depend on choices about all three steps. Most of these choices are made implicitly, in models whose assumptions are rarely visible to readers of the resulting estimate.
4. What does the deal actually commit to?
The deal's environmental architecture is concentrated in its Trade and Sustainable Development chapter. It establishes structured dialogues on forestry, biodiversity, and sustainable agriculture. It commits the parties to "effectively implement" the international environmental agreements which they signed up for. For supporters of the deal, this is the language that turns the agreement into a lever for forest protection. For critics, it commits to nothing in particular.
The question that separates the two readings is enforcement.
The dispute mechanism. The TSD chapter is subject to a state-to-state dispute settlement procedure, an architecture the Commission upgraded in 2022. The upgraded procedure allows panels of experts to issue findings, and provides for the suspension of trade benefits as a last-resort remedy. This is genuinely stronger than what came before. It is also, by the agreement's design, slow, narrowly scoped, and procedurally constrained. ClientEarth's legal analysis of the deal concluded that the dispute mechanism, as drafted, "cannot impose sanctions for non-compliance" with the substantive environmental commitments.
The credible-evidence threshold. A reciprocity provision built into the safeguard regulation that accompanies the deal requires the Commission to "initiate an investigation and adopt safeguard measures where there is credible evidence" that imports do not meet equivalent standards. The Parliament's addition of this clause was the political price of its support. Whether the clause functions in practice depends on how "credible evidence" is interpreted, and on whether the Commission chooses to initiate investigations on environmental grounds or waits for member states or industry to file complaints. Hard Ground #3 noted that the 2024 and 2025 Brazilian audit findings on hormone-treated beef were credible evidence by any reasonable standard. No safeguard investigation followed.
The structured dialogues. The deal's environmental commitments are largely operationalised through structured dialogues — periodic meetings between EU and Mercosur officials on forestry, biodiversity, and related topics. These dialogues produce statements, joint positions, and occasional commitments. They do not produce enforceable obligations.
The deal's environmental architecture is real, but it is designed for cooperation, not for sanction. Confident claims that the deal "includes meaningful protections" depend on a definition of "meaningful" that the deal itself does not provide.
5. On what timescale?
Even if we had a settled counterfactual (we don't, per section one), a settled regulatory environment (we don't, per section two), a settled tariff-to-land-use chain (we don't, per section three), and enforceable commitments (we don't, per section four), the question of what the deal does to the forest still depends on the time horizon on which the question is asked.
The conversation around the deal almost entirely operates in the short horizon. The deal almost entirely operates in the long one. Confident claims about the deal's deforestation effect are usually answering a question — what happens over the next few years — that is not the question the deal is structured to answer.
A six-year estimate captures the direct trade-flow response to the deal's tariff changes. A thirty-year estimate would have to engage with the structural reorganisation of Mercosur agriculture that the deal's market signals will encourage over time and the interactions with variables — climate, infrastructure, regional demand — that are themselves changing over the same period.
The tariff reductions are phased in over ten years for most products and up to fifteen for sensitive ones. The infrastructure investments the deal's market signals encourage — abattoirs, soybean crushing facilities, port capacity, road networks — have payback periods measured in decades. The land-use decisions those investments lock in have consequences over generations of farmers, decades of land use, and the centuries-long question of whether the forest that did or did not exist where they now stand can ever come back.
One thing outside the questions
The five questions above are the questions the conversation has at least tried to ask. There is also a category of questions the conversation has barely engaged with at all. An example is the change in the Brazil-China agricultural trade dynamic.
In 2018, the United States imposed tariffs on Chinese goods. Beijing retaliated by shifting agricultural purchases away from the American Midwest towards Brazil. The structural reorganisation of Brazilian agriculture over the past seven years owes more to that shift than to anything the EU has done or has yet to do.
This is one example. There are others. The growth of Brazilian domestic beef consumption. Currency dynamics that drive frontier expansion more directly than tariff changes. Infrastructure decisions made in Brasília with rising Chinese co-financing. The slow shift in Chinese ESG standards.
Each is documented in the specialist literature. None has crossed into the mainstream conversation about the deal in any sustained way. The pattern of what gets discussed and what does not is itself worth examining.
If, as appears likely, the largest force shaping Mercosur land use over the next decade continues to be the rearrangement of the world's largest soybean trade lane, the EU-Mercosur deal is a smaller piece of the deforestation question than the conversation around it suggests. Not unimportant — just smaller.
At the end of the day, if we care about the forest, we cannot afford to keep looking only at one tree.
A note on process: see my AI Use Policy for how I work with AI in producing this website's content.