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

TL;DR

I forecast 1.99 million metric tonnes of cocoa bean production in Côte d’Ivoire in calendar year 2026. My 80% interval is 1.76 million to 2.20 million tonnes, with a median of 2.00 million tonnes. Strong 2025/26 arrivals lift the first eight months of 2026, while weak 2026/27 pod counts and El Niño pressure cut the September-December slice.

Context

The hard part is not the unit. It is the calendar. FAOSTAT defines the historical series as cocoa beans, production, the amount produced in the year, in tonnes, with a 1961-2024 coverage window and a February 25, 2026 data vintage (OWID/FAOSTAT cocoa-bean production). ICCO is the preferred resolution source, but its May 29, 2026 public bulletin reports the world cocoa balance on an October-September cocoa-year basis, not a current country-level 2026 calendar-year figure for Côte d’Ivoire (ICCO May 2026 QBCS).

Côte d’Ivoire also changed the commercial calendar in 2026. The main campaign now runs from September 1 to February 28, and the small campaign from March 1 to August 31 (Côte d’Ivoire government, February 27, 2026). So calendar 2026 is a blend: the January-August tail of the strong 2025/26 crop plus the September-December start of the weaker-looking 2026/27 main crop.

Evidence

The historical base rate comes from FAOSTAT through OWID. The series has 64 annual observations, covers 1961-2024, uses tonnes, and was retrieved by OWID from FAO Crops and livestock products on February 25, 2026 (OWID/FAOSTAT metadata). Values below are tonnes, rounded to the nearest tonne where needed, and all rows are from the same FAOSTAT/OWID series (OWID/FAOSTAT cocoa-bean production).

The recent regime is the right reference class. Côte d’Ivoire averaged 2.195 million tonnes in 2017-2022, then fell to 1.822 million tonnes in 2023 and 1.890 million tonnes in 2024, a 22.7% drop from 2022 to 2023 followed by only a 3.7% recovery in 2024 (OWID/FAOSTAT cocoa-bean production). The ten-year 2015-2024 mean is 2.031 million tonnes, with annual volatility large enough that a tight forecast would be false precision (OWID/FAOSTAT cocoa-bean production).

Crop-year sources show why the recovery is real but incomplete. USDA/FAS estimated Côte d’Ivoire cocoa bean production at 2.300 million tonnes in MY 2022/23, 1.760 million tonnes in MY 2023/24, and 1.750 million tonnes in MY 2024/25; the same January 7, 2026 report put the MY 2024/25 mid-crop at 500,000 tonnes versus a ten-year average of 550,000 tonnes and flagged swollen shoot disease, dry spells, mold, and weak fermentation as constraints (USDA/FAS GAIN, January 7, 2026). ICCO’s May 29, 2026 update put world 2024/25 production at 4.723 million tonnes, world grindings at 4.628 million tonnes, and the global surplus at 48,000 tonnes, which confirms a move away from the 2023/24 crisis but does not give a clean 2026 Côte d’Ivoire calendar-year number (ICCO May 2026 QBCS).

The best current-season signal is port flow plus the regulator’s own forecast. On May 21, 2026, the Coffee and Cocoa Council said 2025/26 production should reach 2.0-2.1 million tonnes, with port arrivals just above 1.7 million tonnes as of May 11 (Reuters via MarketScreener, May 21, 2026). By July 12, 2026, ADMIS reported cumulative 2025/26 arrivals of 1.956 million tonnes, versus 1.640 million tonnes a year earlier and a five-year average of 1.923 million tonnes (ADMIS, July 14, 2026). Reuters also reported that about 200,000 tonnes could still arrive from July through the end of August (Reuters via Business Recorder, July 12, 2026). I read those data as a final 2025/26 crop near 2.1 million tonnes, not as a return to the 2022 record.

There is a stock and timing caveat. The government said on January 20, 2026 that about 123,000 tonnes of unsold cocoa were still in producing zones, and on March 4, 2026 said more than 1.5 million tonnes had already been paid to producers at the 2,800 FCFA/kg main-crop price (Côte d’Ivoire government, January 20, 2026; Côte d’Ivoire government, March 4, 2026). That means arrivals in early 2026 partly reflect the release of late-2025 beans and marketing bottlenecks. I do not treat arrivals as a pure biological-production measure.

The 2026/27 signal is weaker. Reuters reported on July 12, 2026 that four pod counters and five major exporters expected the 2026/27 main crop, which runs September-February, at 1.35-1.45 million tonnes, down from about 1.6 million tonnes this season; the same report said late-June surveys showed more than 20% flower and cherelle mortality and rising black pod risk after heavy rains (Reuters via Business Recorder, July 12, 2026). IBEX had already reported in June that early pod counts pointed to a 2026/27 Côte d’Ivoire crop of 1.7-1.8 million tonnes before any extra El Niño damage (IBEX Cocoa Market Review, June 8, 2026). On July 13, 2026, Reuters said below-average rain helped dry soils, but farmers still needed more sun and expected clearer crop-size information by late August (Reuters via MarketScreener, July 13, 2026).

The climate signal pushes the downside more than the upside. NOAA’s July 9, 2026 ENSO discussion said El Niño would strengthen through year-end, with a 97% chance of persisting through early spring 2027 and an 81% chance of a very strong event in October-December 2026 (NOAA CPC ENSO Diagnostic Discussion, July 9, 2026). IRI’s June 2026 update put the weekly Niño 3.4 index at +1.7 °C centered on June 17 and said the model ensemble strongly favored an intensifying El Niño through the forecast period (IRI ENSO Quick Look, June 2026). El Niño is not a deterministic Côte d’Ivoire crop model, but it matches the July field reports: too much rain and cloud now, then risk of disrupted rainfall later.

The low official anchor is real but stale. A later DPBEP annex gives cocoa production assumptions, in thousand tonnes, of 2,358.8 in 2022, 1,822.3 in 2023, 1,890.4 in 2024, 1,710.0 in 2025, 1,800.0 in 2026, 1,850.0 in 2027, and 1,900.0 in 2028 (DGBF Annexe 3 DPBEP 2026-2028). I give that 2026 value some weight because it is an official calendar-year-style number. I give it less weight than live arrivals because the July 2026 flow data already show the 2025/26 season is much stronger than the annex implied.

My calendar bridge is this:

$$CY_{2026} \approx (Crop_{25/26} - OctDec_{2025} - StockAdj) + q \times Main_{26/27}$$

Here $Crop_{25/26}$ is the 2025/26 crop-year total, $OctDec_{2025}$ is the part already delivered before 2026, $StockAdj$ is the part of early-2026 arrivals that looks like old unsold stock rather than new 2026 production, $Main_{26/27}$ is the September-February main crop, and $q$ is the share of that main crop produced or commercialized by December 31. Reuters-reported arrivals were 1.028 million tonnes by December 28, 2025, from the October 1 season start (CocoaIntel reprint of Reuters, December 29, 2025). Using a 2025/26 crop around 2.12 million tonnes, subtracting about 1.03 million tonnes already in late 2025 and a modest stock adjustment, then adding roughly two-thirds of a 1.40 million tonne 2026/27 main crop gives a strict-calendar center near 2.00 million tonnes.

For the final distribution, I used a mixture rather than a single normal: 60% on the strict calendar bridge centered at 2.00 million tonnes with 130,000 tonnes standard deviation; 15% on a high source-convention or crop-year-proxy case centered at 2.13 million tonnes with 110,000 tonnes standard deviation; 15% on a low official/stock-adjusted case centered at 1.80 million tonnes with 140,000 tonnes standard deviation; and 10% on broad weather and measurement tails centered at 1.98 million tonnes with 280,000 tonnes standard deviation. That gives a mean of 1.988 million tonnes, a median of 1.996 million tonnes, a 10th-90th percentile range of 1.761-2.199 million tonnes, and a 5th-95th percentile range of 1.685-2.256 million tonnes.

What's non-obvious

The obvious read is to take the CCC’s 2.0-2.1 million tonne 2025/26 forecast as the answer. That is too high for a strict calendar-year interpretation because much of the 2025/26 main crop arrived in October-December 2025, while calendar 2026 includes September-December 2026 from the weaker 2026/27 crop (Reuters via MarketScreener, May 21, 2026; Reuters via Business Recorder, July 12, 2026).

The other trap is treating port arrivals as production. The January 2026 government statement about 123,000 tonnes of unsold beans and the March 2026 statement about more than 1.5 million tonnes already paid show that arrivals were partly a marketing-flow series, not a clean harvest-date series (Côte d’Ivoire government, January 20, 2026; Côte d’Ivoire government, March 4, 2026). That matters because the resolving source may count commercialized output, biological output, or a crop-year proxy.

Limitations

No authoritative 2026 calendar-year production figure exists yet. ICCO is preferred, but the public ICCO release available now is a cocoa-year world-balance update, while FAOSTAT’s annual 2026 value is not expected until a later annual release (ICCO May 2026 QBCS; OWID/FAOSTAT metadata).

The biggest live uncertainty is the September-December 2026 start of the 2026/27 crop. July pod counts can improve or worsen by late August, and El Niño effects depend on local rainfall, sunshine, disease treatment, and farm maintenance (Reuters via MarketScreener, July 13, 2026; NOAA CPC ENSO Diagnostic Discussion, July 9, 2026).

The biggest measurement uncertainty is source convention. If the resolver ultimately uses the 2025/26 crop year as a proxy, the answer should land closer to 2.1 million tonnes. If it uses a strict harvest-date calendar year and subtracts late-2025 stock released in early 2026, the answer can land closer to 1.8-1.9 million tonnes.