How many tonnes of cocoa beans will the Ivory Coast produce in 2026?

Côte d’Ivoire cocoa production in 2026

Executive Summary

The forecast median is 1970000 metric tonnes of cocoa beans, with a central 80% range of 1760000–2190000 metric tonnes. This covers production during January 1–December 31, 2026, not a cocoa marketing season spanning two years.

The median is the middle outcome. The distribution assigns a 10% chance to production at or below the lower endpoint and a 10% chance above the upper endpoint; these are not hard bounds.

The central judgment is that strong early-2026 production partly offsets a weaker, delayed incoming harvest. The weak outlook for the following full crop season should not be substituted for the requested calendar-year total. Uncertainty remains substantial because commercial deliveries, physical production and official year labels do not align perfectly.

Key drivers

Strong production already recorded earlier in the year

Côte d’Ivoire’s national statistics agency, ANStat, published a summary on November 6 reporting 879,818.7 tonnes during January–June 2026, up 58.6% year on year. This is the strongest current-year production observation supporting the forecast. It prevents the weak incoming-season outlook from dominating the entire calendar-year estimate. The underlying monthly returns were not independently verified.

A weak and delayed incoming harvest

Reuters’ November industry assessment described a deteriorating crop outlook, while its August reporting identified an eight-to-ten-week delay in the incoming main harvest. The key remaining physical uncertainty is how much of that crop is harvested before December 31. Delivery delays also matter for interpreting the evidence, but they need not imply an equal loss in production.

The following season’s outlook includes production in 2027. Applying its entire decline to calendar 2026 would therefore overstate the effect on this question.

Calendar boundaries materially affect the answer

The International Cocoa Organization (ICCO) normally reports production for October–September cocoa years. On February 27, 2026, Côte d’Ivoire announced a local marketing-calendar change: the main campaign would run September–February, and the intermediate campaign March–August.

This transition creates a risk of counting September twice when combining reports. FAOSTAT, the UN Food and Agriculture Organization’s statistics database, introduces another uncertainty: its methodology assigns crops to the year in which the harvest, or most of it, occurs, rather than guaranteeing a literal January–December aggregation.

Port arrivals are useful, but are not production

Port arrivals are beans delivered to ports. They provide timely evidence, but storage, residual stocks and delivery delays can separate them from newly harvested output. Official reporting described state purchases of residual cocoa stocks during the completed campaign.

The forecast therefore uses arrivals to reconstruct the year while allowing for differences between commercial flows and production. It does not apply a fixed percentage discount from arrivals to harvest: the research did not establish a defensible adjustment of that kind.

Scenarios

The distribution uses the following judgmental components. Their weights describe how the forecast accommodates accounting and reporting possibilities; they are not measured historical frequencies.

Implications

• Annual output and near-term availability can tell different stories. An improvement over the latest national annual total can coexist with weak late-year deliveries. Procurement decisions should distinguish beans produced from beans available at ports.
• Planning should use the distribution, not just its median. The remaining uncertainty includes both harvest performance and the eventual reported statistic. A single-number budget would conceal these different risks.
• Keep calendar-year and crop-year reporting separate. A weak incoming-season headline does not by itself establish a similarly weak 2026 calendar total. Conversely, a stronger preceding campaign is not wholly attributable to 2026.
• Statistical risk persists after the harvest ends. Source selection, year attribution and permitted revisions may still change the resolving value after December 31.

Signs to watch

The research did not establish exact publication dates for forthcoming national, ICCO or FAOSTAT production updates; none is assumed here.

Research and methodology

The most consequential sources were:

• ANStat production publications. The June 2026 dashboard summary supplied the first-half observation. The December 2025 statistical bulletin supplied recent annual comparisons. Its 2024 total closely matches FAOSTAT’s corresponding value, supporting—but not proving—the relevance of national statistics to eventual resolution.
• Reuters harvest and delivery reporting. The November 4 industry assessment supplied completed-campaign arrivals and expectations for remaining deliveries. Its survey covered 17 industry participants, not a random national farm sample. The August 26 report supplied the prior-year quarter comparison and harvest-delay evidence.
• ADM Investor Services’ arrivals commentary. The November 2 report supplied recent weekly and cumulative arrivals. Its cumulative opening date was not explicit, so the calendar mapping remains an assumption rather than a verified monthly series.
• Official marketing-calendar announcement. The government’s February 27 announcement established the changed campaign boundaries.
• FAOSTAT history and methodology. The FAOSTAT-derived series available through Our World in Data, covering 1961–2024, provided historical scale and volatility checks. FAO’s production methodology established the year-attribution caveat.
• ICCO reporting definitions and availability. ICCO’s statistical definitions established the normal crop-year basis. Its August 2026 bulletin stated that relevant production estimates were temporarily withheld, so it did not provide an observed answer.

The quantitative work reconstructed calendar-year commercial flows by removing the late-2025 portion of the preceding campaign and adding the relevant late-2026 flows, while avoiding overlapping reporting windows. Official first-half production was used as an anchor and consistency check, not added to overlapping arrivals. A main-harvest accounting cross-check did not fully reconcile with the delivery reconstruction, supporting a wider distribution. Reporting alternatives and a broad error allowance were then incorporated into the final probability distribution.

Limitations

• No official later-year monthly production total or inventory reconciliation was obtained. Those data would most directly improve the forecast. The first-half figure was available through ANStat’s official publication summary, but its underlying monthly returns were not independently verified.
• The source mapping remains incomplete. A close historical match does not establish that ANStat directly feeds FAOSTAT or that their revision and year-attribution practices are identical. Transitional campaign coverage and cumulative-arrivals starting dates also remain imperfectly documented.
• The uncertainty allowances are judgmental. They are not measured error rates from a validated arrivals-to-production model.

The distribution assumes a numeric resolution. Under the question’s rules, it will instead be annulled if no authoritative data are available by June 30, 2029.