Brazilian soy exports and deforestation

Brazil is the world’s largest soy producer and exporter. In 2023, it produced almost 152 million tonnes of soy, of which 127.3 million tonnes was exported as 80% raw soybeans, 18% soybean cake and 2% soybean oil. Soy was Brazil’s second largest export, accounting for 16% of total exports in 2023 and generating an annual revenue of almost US$53 billion.

Since 2019, the price of soy has risen due to demand for soy-based animal feed from China and Russia’s war in Ukraine. Soy prices in 2024 have stabilised but remain high, potentially incentivising further deforestation and conversion to expand soy plantations.

Trase data shows the amount of deforestation and conversion linked to soy production increased from 635,000 hectares (ha) in 2020 to 794,000 ha in 2022, while the total area of soy planted increased from 37.2 million hectares (Mha) in 2020 to 41.2 Mha in 2022. Despite this increase of 4 Mha in the total area of soy plantations (2020–2022), Brazil saw a decrease of 14 million tonnes in soy production between 2021 and 2022. This decline was due to droughts caused by La Niña in the southern region of the country.

Soy can also be an indirect driver by expanding onto cattle pastures that were the original driver of deforestation and conversion. Of the 39.8 Mha of soy planted in 2023, 6.7% (2.6 Mha) was pasture in 2015.

The Cerrado and Pampas are active hotspots of deforestation and conversion

Soy plantations have expanded most in the Cerrado and in 2022 were associated with 375,000 ha of deforestation and conversion – an area more than twice the size of the city of São Paulo. The Pampas has also experienced high rates of conversion of natural vegetation to soy. In 2022, 250,800 ha of soy was harvested in areas recently deforested and converted in the Pampas.

For comparison, soy production in the Amazon in 2022 was linked to 117,000 ha of deforestation – much lower than the Cerrado and Pampas biomes.

Although only 6.3% of soy deforestation in 2022 occurred in the Atlantic Rainforest (50,000 ha), it is of particular concern as it has been illegal to clear native Atlantic forests since 2006.

Please note that the latest version of the MapBiomas database used by Trase to calculate deforestation and conversion in the Pampas, Pantanal, Caatinga and Atlantic Forest accounts more accurately for the loss of other wooded lands and natural grasslands than previous versions, resulting in higher estimates of the total area of soy deforestation and conversion. This is particularly significant in the Pampas, for which total deforestation and conversion for soy in 2008–2023 changes from 363,800 ha to 508,800 ha. For the whole of Brazil over the same period, the estimate changes from 2.07 million ha to 2.56 million ha.

The EU deforestation regulation creates challenges and opportunities for Brazilian soy producers

The EU deforestation regulation (EUDR) entered into force in 2023 with binding obligations from 30 December 2025. It will require mandatory due diligence on imports of certain agricultural commodities, including soy from Brazil, to prevent products grown on land converted or deforested after the 31 December 2020 cut-off date. The regulation establishes a benchmarking system to classify countries as low, standard, or high risk, to both facilitate simplified due diligence processes by operators sourcing from low-risk countries and enable competent authorities to effectively target enforcement.

Trase data shows that in 2022, just 370 of a total of 2,525 soy-producing municipalities accounted for 95% of deforestation and conversion for soy in Brazil (between 2017 and 2022). These municipalities represented 58% of Brazil’s soy production in 2022 (70 million tonnes) and 52% of exports (51 million tonnes). Of the 370 municipalities, 123 are located in the Matopiba region, a major hotspot accounting for 38% of the total deforestation and conversion for soy in 2022. In the state of Rio Grande do Sul (Pampas biome), 75 municipalities that produced soy accounted for 33% of the total deforestation and conversion linked to the commodity in 2022.

This means that the majority of soy exports in 2022 are associated with less than 5% of Brazil’s soy deforestation and conversion, highlighting that the risks of non-compliance with the EUDR is concentrated in certain regions. This demonstrates the value of the EUDR benchmarking system in classifying the risk of subnational production regions, particularly in large countries such as Brazil where deforestation is concentrated in specific regions. Similarly, the use of subnational risk assessment by competent authorities to target checks could focus enforcement where it is most needed.

Deforestation exposure of export markets

China continues to be the most exposed market to deforestation from its soy imports, followed by Brazil’s domestic market and imports by the European Union (EU).

Trase’s ability to link all Brazilian soy exports to municipalities of production based on publicly available data is a challenge for the 2020–2022 period due primarily to missing information in trade data, and a lack of information on trader silo and crushing facility ownership. As a result, 15–18% of soy traded in 2020–2022 (approximately 20 million tonnes) cannot be linked to a specific municipality of soy production (unknown municipality). To account for this, Trase’s Brazilian soy supply chain data now includes deforestation exposure of trading companies and markets associated with their sourcing from unknown municipalities.

To do this, Trase first assigns deforestation exposure to companies and importing markets according to their sourcing from known soy-producing municipalities based on the municipality-specific soy deforestation per tonne. Second, the remaining, unaccounted-for soy deforestation is allocated to companies and import market trade flows originating from unknown municipalities. Similar to the first step, deforestation per tonne is calculated by dividing the remaining soy deforestation by the total production originating from unknown municipalities, and assigned to companies and markets based on their traded volume sourced from unknown municipalities. This prevents importing countries and traders with a greater volume of soy of unknown origin from appearing to have a lower deforestation exposure. It also allows for the allocation of all of Brazil’s soy deforestation to domestic and international supply chains.

For some importing countries, there is a significant difference in the amount of soy and deforestation that can be linked to specific soy-producing municipalities. For example, in 2022, 97% of China’s soy imports and 96% of its associated deforestation exposure could be linked to soy-producing municipalities, whereas for the EU, only 58% of imports and 55% of deforestation exposure could be linked. Given such differences, it is important for users to understand that importing countries or companies with a higher degree of unknown sourcing will have a lower proportion of their deforestation exposure estimate that is specific to their sourcing pattern, as their unknown sourcing deforestation exposure is the average deforestation per tonne of all unknown trade flows.

Traders with the highest exposure to soy deforestation and conversion

The large, established traders – Bunge, COFCO and Cargill – are the most exposed to deforestation and conversion in their soy supply chain, respectively accounting for 77,700 ha (9.6%), 67,800 ha (9%), and 55,100 ha (8%) in 2022. In 2020–2022, the exposure of Bunge and Cargill remained stable, while COFCO’s exposure increased significantly from 27,100 ha in 2020. This is due to COFCO’s sourcing from the Pampas and, as noted above, the enhanced ability of Mapbiomas to detect the conversion of other wooded lands and natural grasslands.

NovaAgri, a Brazil agribusiness company and subsidiary of Toyota Tsusho Corporation, was the fourth most exposed trader to deforestation and conversion in 2022, despite only being seventh for soy exports.

As discussed above for export markets, Trase’s Brazilian soy supply chain data now includes deforestation exposure to soy sourced from unknown production regions. It is important that users understand the implications of this when interpreting the data on deforestation exposure of trading companies.

Zero-deforestation commitments outside the Amazon fall short

In 2022, 59 million tonnes of soy (79% of the total volume) were traded by companies with a zero-deforestation commitment (ZDC).

The Amazon Soy Moratorium, a voluntary commitment signed by over 25 companies, is the most effective ZDC in Brazil. In 2022, 92% of soy from the region (7 million tonnes) was sourced by signatory companies. Areas covered by the moratorium have relative rates of direct deforestation for soy that are much lower than other soy production regions. The deforestation intensity of soy production in the Amazon is 3.2 hectares per 1,000 tonnes in 2022 in areas covered by the moratorium, compared to 5 hectares per 1,000 tonnes for production not covered by any ZDC. This shows the importance of the moratorium in promoting the sustainable production of soy in the Amazon and the need to resist attempts to weaken it.

Outside the Amazon, 73% of total soy production (52 million tonnes) was traded by companies with a ZDC. Despite this coverage, deforestation and conversion for soy remains high in biomes such as the Cerrado and Pampas, which means that companies’ ZDCs appear to be much less effective at preventing deforestation.

Carbon emissions from soy deforestation and conversion

Deforestation and conversion linked to the expansion of soy production in 2022 resulted in the release of 121 million tonnes of CO₂ – 9% of the country's total land-use change annual emissions. Soy replaced three times as much native vegetation in the Cerrado as the Amazon, but emissions from land clearance were only 42% higher in the Cerrado (49 million tonnes of CO₂) when compared to the Amazon biome (28 million tonnes of CO₂).

The authors thank the researchers and data scientists who have contributed to this analysis: Harry Biddle, Florian Gollnow, Michael Lathuillière, Nicolás Martín, Carina Mueller, Vivian Ribeiro and Clément Suavet.

Explore our new data on the Brazil soy supply chain

A detailed explanation of Trase’s methodology is available at: Trase. (2025). SEI-PCS Brazil soy v2.6 supply chain map: Data sources and methods. Trase. https://doi.org/10.48650/X24R-YK29

To reference this article, use the following citation: Pereira, O., & Bernasconi, P. (2025). Brazilian soy exports and deforestation. Trase. https://doi.org/10.48650/Q48G-MJ07