Ethiopia is arabica's birthplace, Africa's largest coffee producer, and the origin where EUDR's data demands collide hardest with how coffee actually reaches the port. Millions of smallholder households grow it, much of it under tree canopy, and a share of exports still moves through marketing channels that were designed to anonymise origin. This page covers the Ethiopian specifics on top of the general coffee guide.
Tier and baseline obligations
Ethiopia defaults to standard risk: full due diligence, documented risk assessment, and a DDS with complete plot geolocation. Holdings are small, overwhelmingly under 4 hectares, so point declarations carry the load, with garden plots often well under half a hectare.
Four production systems, four screening stories
Ethiopian coffee is conventionally split into garden, semi-forest, forest and plantation systems. The distinction matters for EUDR more than anywhere else:
- Garden coffee (the majority): small plots around homesteads, mixed with enset and other crops. Screening is straightforward; the work is collection volume.
- Semi-forest and forest coffee: coffee harvested under thinned or intact natural canopy in the southwest. These plots read as forest in tree-cover datasets, because they are forest. That is not a compliance problem in itself: the test is conversion after 31 December 2020, not the presence of trees. But it makes single-layer screening useless and convergence of evidence essential, and it puts a premium on honest plot boundaries rather than village centroids.
- Plantation coffee: a small share of larger private and state farms, above 4 hectares, needing polygons.
The screening risk to actually manage is gradual canopy thinning: intensifying a forest coffee stand can shade into degradation, and new clearings for coffee at forest margins in the southwest do occur. Plots flagged near the cutoff need production dates and imagery history, not assurances.
Traceability: the ECX legacy and what replaced it
For years most Ethiopian coffee had to trade through the commodity exchange, which graded lots by cupping profile and region while deliberately breaking the link to individual suppliers. Reforms since 2017 reopened direct and vertically integrated export paths, and cooperative unions market members' coffee with grower records attached. The consequence for EUDR buyers: traceability depends heavily on the channel. A union or single-estate lot can come with grower lists and plot points; an exchange-sourced blend may trace only to a washing station's catchment. For the latter, the workable pattern is supply-shed mapping: the station maps its delivering farmers once, using a field-collection campaign, and every subsequent lot reuses that plot file with fresh production dates. Exporters increasingly offer exactly this; contract for it explicitly.
Practical package per lot
- Channel identification: union, estate, or station-blend, because it sets what data can exist.
- Plot file: points with areas and grower identifiers for the supplying farmers, polygons for any plantation blocks.
- Production dates for the harvest season in the lot.
- Your own screening run, with forest-coffee plots assessed on conversion evidence rather than tree cover alone.
- DDS filed in TRACES, evidence retained five years.
For the same smallholder mechanics with stronger institutional data, compare Colombia; for the procedural opposite, the low-risk track in Vietnam.