Compliance Checker Methods
Development of EU member state risk profiles
Identification of deforestation-risk countries
AidEnvironment identified an initial longlist of countries with potential deforestation risk to include in the Compliance Checker, based on a range of criteria, including the presence of intact forests, net forest loss, and relevant commodity production. The criteria are a combination of the FAO definitions of forest, and a selection of risk assessment criteria under Article 10 of the EUDR:
- Country has presence of ‘forests’ in its territory and produce relevant commodity or product, with ‘forests’ as defined by the FAO and the EUDR. The forest definition will also include parts of the Cerrado and Chaco biomes in Latin America, as well as wooded savannas and woodlands in Africa, based on FAO’s definition of ‘open forests’. Moreover, priority is given to countries with ‘primary forests’, which are a fundamental priority for conservation under FAO definitions and the EUDR. Finally, countries with ‘secondary forests’ (often degraded) are also included, which are ‘forests’ (by definition), and can also still contain high canopy cover (>10%) and biodiversity.
- The prevalence of deforestation or forest degradation in the country, region, and area of production of the relevant commodity or product.
AidEnvironment has used the FAO’s Global Forest Resources Assessment (FRA) 2025 to prepare a long list of countries that would fall under these definitions. The report included 236 countries and territories in the FRA 2025, based on the list used by the United Nations Statistics Division. For each country, forest area (2025) and forest loss (2015–2025) were assessed. Based on the criteria ‘presence of forests’ and ‘prevalence of deforestation’, 83 countries were selected in the long list. To assess relevant commodity production in this long list of countries, we have used FAOSTAT 2023 production data. This would guarantee that we only include a country for commodity-linked deforestation if the commodity is also actually produced in that country. Moreover, we have used European trade statistics (Access2Markets) for assessing 2024 trading links between the producing country and the EU. The evaluation of actual commodity production and existing trading linkages with the EU27 countries, reduced the longlist of 83 countries to 62 countries with deforestation risk linked to relevant commodity production. As a threshold for forest cover, we have set (an arbitrary) 200,000 hectares, resulting in excluding two more countries (Singapore and Israel), leaving us with 60 remaining countries.
Still, some challenges remain in the identification of high-risk countries based on the aforementioned criteria. Countries such as Australia, Ghana, Philippines, Viet Nam, and India would not be included based on the criteria since the country self-reports no net forest loss. Instead, these countries report an increase in forest area. However, based on more detailed deforestation-risk analysis, AidEnvironment is aware of commodity-linked forest loss in Australia, e.g., for beef. Ghana is historically linked to high deforestation-risk linked to cocoa expansion. Agricultural expansion as well as logging have had a significant impact on the deforestation in Philippines from 1980 – 2020. Rubber and coffee are attributed to the forest loss in Viet Nam, as well as recent palm oil expansion in India. Apart from their link to deforestation, these five countries export large amounts of these commodities to the EU, a reason to include them in the list of countries. While the same reasoning may apply to other countries, AidEnvironment did not do a full risk assessment for each country to understand the drivers of deforestation.
Finally, for the category of wood products, no comparable production data (e.g. FAOSTAT) is available. However, we employ EU trade statistics data to analyze timber exports from countries experiencing deforestation.
| Country | Forest area (1,000 ha) (2025) | Forest Loss (%) (2015–2025) | Commodities' production (2023)** |
|---|---|---|---|
| Angola | 63,262 | -0.77 | (S), (B), (L), (P), (C), (CF) |
| Argentina | 46,598 | -0.39 | (S), (B), (L) |
| Bangladesh | 1,873 | -0.05 | (S), (B), (L), (R) |
| Belize | 1,277 | -0.85 | (S), (B), (L), (C), (CF) |
| Bolivia | 54,370 | -0.42 | (S), (B), (L), (C), (CF), (R) |
| Brazil | 486,087 | -0.59 | (S), (B), (L), (C), (P), (R), (CF) |
| Burkina Faso | 3,237 | -3.10 | (S), (B), (L) |
| Cambodia | 6,333 | -3.29 | (S), (B), (L), (CF), (R) |
| Cameroon | 19,143 | -0.74 | (S), (B), (L), (C), (P), (CF), (R) |
| Central African Republic | 45,095 | -0.10 | (B), (L), (C), (P), (CF), (R) |
| Chile | 1,442 | -0.41 | (B), (L) |
| Colombia | 59,457 | -0.25 | (S), (B), (L), (C), (P), (CF), (R) |
| Congo | 21,854 | -0.03 | (B), (L), (C), (P), (CF), (R) |
| Côte d'Ivoire | 3,774 | -1.63 | (S), (B), (L), (C), (P), (CF), (R) |
| DRC | 139,189 | -0.20 | (S), (B), (L), (C), (P), (CF), (R) |
| Ecuador | 12,310 | -0.40 | (S), (B), (L), (C), (P), (CF), (R) |
| Equatorial Guinea | 2,407 | -0.34 | (B), (L), (C), (P), (CF) |
| Ethiopia | 26,747 | -0.01 | (S), (B), (L), (CF) |
| Gabon | 23,555 | -0.03 | (S), (B), (L), (C), (P), (CF), (R) |
| Guatemala | 3,536 | -0.38 | (S), (B), (L), (C), (P), (CF), (R) |
| Guinea | 4,857 | -1.28 | (B), (L), (C), (P), (R), (CF) |
| Guinea-Bissau | 2,100 | -0.02 | (B), (L), (P), (R) |
| Guyana | 18,377 | -0.05 | (B), (L), (C), (CF) |
| Honduras | 5,861 | -1.07 | (S), (B), (L), (C), (P), (CF) |
| Indonesia | 95,969 | -0.10 | (S), (B), (L), (C), (P), (R), (CF) |
| Iraq | 693 | -0.91 | (S), (B), (L) |
| Japan | 24,908 | -0.01 | (S), (B), (L) |
| Laos | 13,036 | -0.25 | (S), (B), (L), (CF), (R) |
| Liberia | 6,327 | -1.06 | (S), (B), (L), (C), (P), (CF), (R) |
| Madagascar | 9,922 | -1.16 | (S), (B), (L), (C), (P), (CF) |
| Malaysia | 18,885 | -0.30 | (B), (L), (C), (P), (CF), (R) |
| Malawi | 2,032 | -1.86 | (B), (S), (L), (CF) |
| Mali | 10,497 | -0.91 | (S), (B), (L) |
| Mexico | 66,266 | -0.19 | (S), (B), (L), (C), (P), (CF), (R) |
| Mozambique | 32,243 | -0.79 | (S), (B), (L), (CF) |
| Myanmar | 27,095 | -1.01 | (S), (B), (L), (R), (CF) |
| Namibia | 8,043 | -0.01 | (B), (L) |
| Nicaragua | 4,781 | -0.45 | (S), (B), (L), (C), (P), (CF) |
| Nigeria | 17,130 | -0.71 | (S), (B), (L), (C), (P), (CF), (R) |
| Panama | 4,615 | -0.07 | (S), (B), (L), (C), (P), (CF) |
| Papua New Guinea | 34,029 | -0.04 | (B), (L), (C), (P), (CF), (R) |
| Paraguay | 14,297 | -1.34 | (S), (B), (L), (P), (CF) |
| Peru | 67,160 | -0.35 | (S), (B), (L), (C), (P), (R), (CF) |
| Republic of Korea | 6,279 | -0.09 | (S), (B), (L) |
| Republic of Moldova | 370 | -0.42 | (S), (B), (L) |
| Senegal | 8,649 | -0.18 | (S), (B), (L), (P) |
| Sierra Leone | 2,436 | -0.78 | (B), (L), (C), (P), (CF) |
| Solomon Islands | 2,514 | -0.03 | (B), (L), (C), (P) |
| Sudan | 21,980 | -0.76 | (B), (L) |
| Surinam | 14,674 | -0.11 | (B), (L), (P) |
| Thailand | 19,647 | -0.21 | (S), (B), (L), (C), (P), (CF), (R) |
| Timor-Leste | 1,054 | -0.11 | (S), (B), (L), (CF), (C), (R) |
| Trinidad and Tobago | 226 | -0.19 | (B), (L), (C), (CF) |
| Tunisia | 687 | -0.25 | (B), (L) |
| Uganda | 2,368 | -1.12 | (S), (B), (L), (C), (CF) |
| United Republic of Tanzania | 43,400 | -1.02 | (S), (B), (L), (P), (CF), (C) |
| Vanuatu | 907 | -0.26 | (B), (L), (C), (CF) |
| Venezuela | 47,088 | -0.17 | (S), (B), (L), (C), (P), (CF) |
| Zambia | 44,874 | -0.19 | (S), (B), (L), (CF) |
| Zimbabwe | 13,766 | -0.38 | (S), (B), (L), (CF) |
Source: AidEnvironment, based on Global Forest Resources Assessment (FRA) 2025, FAOSTAT 2023 production data, and 2024 EU trade statistics. S=Soy, B=Beef, L=Leather, P=Palm oil, CF=Coffee, C=Cocoa, R=Rubber. For Wood, no comparable (FAO) production data is available.
Identification of exporters and importers
AidEnvironment identifies relevant exporters and importing operators from (paid) shipment data based on availability. Shipment data is advantageous for identifying key value chain actors because it is based on recent and hard evidence-based data, not on modelling. However, shipment data also has limitations.
Limitations of shipment data
1) Funding dependence. First, shipment data is paid data, which makes AidEnvironment dependent on funding to acquire it. At the same time, this evidence-based data is necessary to improve transparency in global commodities supply chains.
2) Coverage varies by provider/country. Second, we depend on which shipment data is available from the data provider. For several countries, no shipment data is available from AidEnvironment’s paid shipment data provider (e.g. on China). There is shipping data available for other countries, such as Brazil, Côte d’Ivoire, Indonesia, Vietnam, and Argentina, but many datasets have their limitations. For instance, Indonesia had no shipment data available between September 2021 and December 2023, as the Indonesian government stopped sharing customs data during that period. Since January 2024, Indonesian shipment data is accessible in our paid shipment provider. Vietnam data does not provide any quantities, only values, and Argentina data only provides exporters' names and no importing companies. For each relevant analysis, we have made explicit what limitations exist regarding available shipment data and what is covered or not covered (see Table below).
| Commodity | HS code product | Description product (main traded products within each commodity group) | Major linked (deforestation-risk) supplying country (in 2023) | Shipment data and EU operators available? |
|---|---|---|---|---|
| Beef & Leather | 0202 | Frozen bovine meat | Brazil | YES, but it requires a filter on the destination port instead of the destination country |
| 4107 | Prepared leather products after tanning or crusting | Brazil | YES | |
| 0201 | Meat of bovine animals, fresh or chilled | Argentina | PARTIALLY, data lacks EU importer names, and the export dataset is so extensive that the data cannot be accessed at once/is very expensive | |
| 4101 | Raw hides and skins of bovine animals | None (top-3 suppliers are UK, Norway, and Switzerland) | NO, no shipment data for the top-3 suppliers | |
| 4104 | Tanned crust hides and skins of bovine animals | Brazil | YES | |
| Soy | 1201 | Soybeans | Brazil | YES, but relatively many unnamed records |
| 2304 | Soybean oilcake (=soybean meal) | Brazil | YES, but relatively many unnamed records | |
| Rubber | 4001 | Natural rubber | Côte d’Ivoire | YES |
| 4011 | New pneumatic tires of rubber | China | NO, no available data for China | |
| Palm oil | 1511 | (Crude) palm oil | Indonesia | YES, but no customs data between Sept 2021 and December 2023. Available again since January 2024 |
| 3823 | Palm oil derivatives – industrial fatty acids, oils and alcohols | Indonesia | YES, but no customs data between Sept 2021 and December 2023. Available again since January 2024 | |
| Cocoa | 1801 | Cocoa beans | Côte d’Ivoire | YES |
| Coffee | 0901 | Coffee beans | Brazil | YES |
Source: AidEnvironment, based on data availability of our shipment data provider
3) It is a sample, not a census. The third limitation when considering shipment data is that it needs to be seen as a sample and not as a dataset covering the entire trading flows between producing and consumer countries, for instance, because relevant exporters and importers might be part of traded volumes under logistical companies or the reported volume by shipment data might represent only part of the total volume traded as explained below.
4) Discrepancies vs. EU trade statistics. Fourth, discrepancies exist between European trade statistics and shipment data. We have been informed by trade data providers that discrepancies can exist, among others, because there are sometimes not-reported records in trade data (there is no more information on this, e.g., under what conditions and for whom this applies). For some databases, for instance linked to soy products from Brazil, we observed many unnamed (“null”) records. For some of the key soy traders, such as Bunge, Cargill, and Louis Dreyfus Commodities (LDC) we observed in available shipment records that when they played the role of ‘carrier’ in the shipments, they simultaneously also played the role of importing operator. Therefore, likely these traders can be assigned to the unknown (“null”) records when they also carry the loads.
5) Non-uniform time spans. Finally, shipment data covers many different time spans, making it challenging to create uniform trade data for a specific period of time. This is also linked to the fact that shipment data availability differs per country and commodity. Therefore, the timelines of the European trade statistics may also differ from the most recent available shipment data.
Units and conversion. AidEnvironment analysis indicates trade volume units where necessary. The most common indication in shipment data records is the weight (in kilograms), but others have data in metric tons (MT) or other units (e.g., cocoa in jute sacs). AidEnvironment analysis converts all these different units in kilograms (KG) or metric tons (MT) to make the comparison easier.
Corporate analysis and company grouping
After downloading raw shipment data, the next step is cleaning the data, which includes grouping exporting and importing companies into company groups. Grouping companies requires doing targeted research on the relevant entities to understand their corporate structure, including on their subsidiaries and possible affiliations, as well as their role in the supply chain (e.g., producer, logistical company, commodity trader, downstream company). For this, AidEnvironment resorted to different strategies, depending on the commodity/sector at stake, and used a combination of different tools and sources, namely:
- Orbis – A Moody’s Analytics comparable data resource on private companies (also covering listed companies) that allows assessment of corporate ownership structures, current subsidiaries and controlling shareholders, among others. Link
- AidEnvironment’s databases on company linkages – Information gathered throughout years of work in different geographies and supply chains. For palm oil, for instance, we have mapped palm oil mills and oil palm concessions, listing company group structures based on notary acts/company-registry profiles.
- U.S. Securities and Exchange Commission (SEC) filings – The U.S. regulator collects financial and operational information of publicly traded domestic and foreign companies, which are required to submit financial statements, periodic reports and other documents. Some of the information made public by the SEC are subsidiary lists of companies (e.g., list of subsidiaries of JBS). Example link
- Companies’ websites – In several cases, company groups disclose on their websites the affiliated companies (e.g., Conceria Priante, part of the JBS group). Link
- Sector-specific working groups or roundtables – For instance, the RSPO and the Leather Working Group gather lists of members/certified suppliers from which it is possible to derive company affiliations. RSPO members · Leather Working Group
- Publicly available online sources – AidEnvironment also executes background research online on companies, using reliable and well-identified sources.
Data used for the development of case studies
For the development of case studies, AidEnvironment relies on the public availability of country-specific data sources. Generally, we rely on our various geospatial datasets developed over many years, typically including cadaster/concession data, company assets, supply chain data, deforestation and fire data, vegetation types (e.g., peatland, wooded savanna), and social data, e.g., on land conflicts and encroachment in protected and indigenous territories. We validate our case studies with local partners in origin producer countries. For a full list of reference data sets used per country, please contact AidEnvironment.
In Brazil, AidEnvironment combines geospatial data of rural cadaster systems (CAR, SIGEF, SNCI); company asset infrastructure (warehouses/silos, palm oil mills, slaughterhouses, tanneries), deforestation (e.g. Prodes, MapBiomas), supply chain data (e.g. animal transit data, the so called GTA), fire alerts (NASA), type of vegetation, locations of indigenous territories (e.g. Funai) and protected areas (e.g. Forest Code) to calculate deforestation exposure risk and legality issues of companies.
Carbon emissions in Brazil
Only for Brazil, AidEnvironment was able to develop a tool to calculate carbon emissions based on native vegetation type and amount of cleared area. For this, we have assessed “above-ground carbon emissions” according to the cleared vegetation type(s). The tons of carbon emission were calculated per type of vegetation times the ratio of molecular weight of carbon dioxide to carbon (44/12) times the number of hectares cleared.
Our sample of direct and indirect suppliers to the largest Brazilian meatpacker companies
AidEnvironment is known to have one of the best datasets on cattle movements combined with farmer cadaster data in Brazil. Currently, we have a sample of 36,441 direct and 103,932 indirect supplying farms to Brazil’s largest meatpackers JBS, Marfrig, and Minerva. AidEnvironment uses this dataset to monitor deforestation and fires within the meatpackers’ supply chains, contributing to Civil Society Organizations campaigns such as Mighty Earth’s Rapid Response and Rainforest Foundation Norway’s leather report, and articles garnering media attention, for instance in The Washington Post and The Guardian.