Cocoa sustainability: promise and outcome

Most of what is published about sustainable cocoa describes programmes: what will be paid, planted, monitored or banned. This page sets each promise beside what somebody later measured, and says what kind of measurement it was.

7 promises checkedSurveys, satellites, studiesMembership is not an outcome

Four columns, kept apart

Promise: What a scheme, a law or a pledge says will happen. A price announced is a promise.
Implementation: Whether it was put in place. Joining a certification scheme belongs here.
Measured outcome: What somebody counted afterwards, on farms or from orbit.
Evidence quality: What kind of counting: a representative survey, satellite mapping, a comparison that cannot separate cause from selection, a model, interviews, or nothing.

Promise against outcome

Each row is supported, with its sources, in the section named below it on this page.

Cocoa sustainability promises set against measured outcomes and the quality of the evidence
Subject	Promise	Implementation	Measured outcome	Evidence quality
Child labour	A 70% reduction in the worst forms of child labour in the cocoa sectors of Côte d'Ivoire and Ghana by 2020.	Not established from any source read.	Not met. In 2018/19, 45% of children in agricultural households in cocoa areas were in child labour in cocoa, and the share in hazardous work had risen over ten years.	Measured: a representative survey
Farmer income	A living income: benchmarks of what a household needs have been published for both countries.	Minimum prices, premiums and a living-income differential on exports.	In Ghana, 9.4% of typical male-headed cocoa households reached the benchmark, on 2016 income data.	Measured: a representative survey
Producer prices	Ghana: at least 70% of the export price to the farmer, by law. Fairtrade: a minimum price and a premium.	Prices are announced each season by the two national bodies.	What farmers actually received, net of deductions, was not measured by any source read.	No measurement found
Certification and income	Better prices and incomes for certified farmers.	Certified farms are concentrated in some regions more than others, and not where poverty is deepest.	Modest and uneven. A review across crops found higher prices and no higher household income; in cocoa, outcomes were mixed in Ghana and Ivory Coast, spending rose 9% under Fairtrade in Côte d'Ivoire, and gains in Indonesia depended on which company ran the scheme.	Weak: comparisons that cannot separate cause from selection
Deforestation	Deforestation-free cocoa, which becomes a European legal requirement from 30 December 2026.	The European rule's main obligations have been postponed twice and had not yet applied when this was written.	Cocoa is tied to over 37% of forest loss in protected areas in Côte d'Ivoire and over 13% in Ghana.	Measured: satellite mapping
Traceability	Cocoa traceable to the farm or plot.	Traders monitor the suppliers they buy from directly.	Traders often buy more than 40% of their supply indirectly, through intermediaries outside that monitoring.	Weak: comparisons that cannot separate cause from selection
Higher yields sparing forest	Intensify existing farms and the forest will not need clearing.	Not assessed here.	A model estimates the forest that would have been spared. A household survey found more clearing where cocoa income was higher.	Modelled, not measured

Child labour

The best-measured subject on this page, because two governments and an industry agreed to a number and a university was paid to check it.

What counts. Not all work by children is child labour: work above the minimum age that does not harm health or schooling is excluded. Hazardous work is one of the worst forms. The survey counted a child as in child labour if under 12 and working, working beyond the hours allowed for their age, or doing hazardous work at any age; chores at home are not counted. (source 1; source 2)
The promise: a 70% reduction in the worst forms of child labour in the two countries' cocoa sectors by 2020. The survey concluded that the target was not met: the share of children in hazardous work rose by 14 percentage points between 2008/09 and 2018/19, while cocoa production rose 62%. (source 2)
The measurement: in 2018/19 about 1.56 million children were in child labour in cocoa production, 45% of children aged 5 to 17 in agricultural households in the cocoa-growing areas: 38% in Côte d'Ivoire and 55% in Ghana. (source 2)
About 1.48 million were exposed to at least one form of hazardous work: clearing land, carrying heavy loads, agro-chemicals, sharp tools, long hours or night work. (source 2)
The denominators matter. These are percentages of children in farming households in cocoa areas, not of all children in either country. (source 2)
Trends are harder than they look. The survey's authors say counts cannot be compared with the earlier rounds because the sampling frames differed, and that prevalence rates are not fully comparable. (source 2)
How the question is asked changes the answer. Against field data collected at harvest, parents under-reported their children's work by at least 60%, and children's own reports matched. (Lichand and Wolf 2025)
What the harm is. Children harvesting in Ghana were observed using sharp tools and pesticides without training or protection, and among children at school it was hazardous work, not farm work as such, that went with repeating a year. (Mull and Kirkhorn 2005; Pirkle et al. 2024)
What would it cost to end. One household model puts the cocoa price premium needed to remove the worst forms at 2.81% in Ghana. It is a model's answer and has not been tried. (Luckstead et al. 2019)

Not shown: Whether monitoring and remediation systems reduce child labour. No evaluation with a measured outcome was retrievable.

Not shown: Anything about forced labour, which is a different thing with different evidence and is not covered by these surveys.

Not shown: The situation since 2019. No later survey of comparable scale was read.

Farmer income and living income

The benchmark, which is a statement of need and not a measurement of income: for rural cocoa areas of Ghana, 1,464 cedis (329 US dollars) a month for two adults and three children at March 2018 costs; for Côte d'Ivoire, 262,056 CFA francs a month in August 2018, updated to 298,983 for June 2022. (source 3; source 4)
The measurement: a survey of 3,045 households found average income from all sources of 2,487 US dollars a year in Ghana and 2,900 in Côte d'Ivoire at 2016 values, of which cocoa supplied 61 and 66%. (source 5)
After inputs and hired labour, cocoa alone left the average household 1,510 dollars in Ghana and 1,908 in Côte d'Ivoire. (source 5)
Set against the benchmark by the survey's own authors, 9.4% of typical male-headed households in Ghana reached it; among those with large landholdings, 44% did. (source 6)
Six household data sets from the two countries give the same picture: 73 to 90% of households below a living income and 30 to 58% below the World Bank's extreme poverty line. Doubling the cocoa price would leave 53 to 65% still below a living income. (van Vliet et al. 2021)
Farm size is the recurring reason. An analysis of existing data sets concluded that a large proportion of cocoa smallholders cannot reach a living income from their present farms, because the farms are small and the means to invest are lacking, and that there is no single remedy. (Waarts et al. 2021)

Not shown: Incomes since the price rise of 2023 to 2025. The household survey is from 2016.

Not shown: An equivalent share for Côte d'Ivoire. The report that gives it was not obtained.

Not shown: That a higher price alone would close the gap. The land figures point the other way for the smallest farms.

Prices: announced, guaranteed, and received

Three different things. Only the first two are documented here.

Ghana announced a producer price of 42,400 cedis a tonne for the 2026/27 season, effective 25 September 2026, and states that it is 71.18% of the realised export value. The cocoa board says a 2026 act guarantees the farmer at least 70%. (source 7)
Côte d'Ivoire's cocoa council showed a guaranteed minimum producer price of 1,200 CFA francs a kilogram on 7 October 2026. (source 8)
Fairtrade's own rules set a minimum price of 3,500 US dollars a tonne for Ghana and 3,200 euros for Côte d'Ivoire at export level, with a premium of 275 dollars or 250 euros, mandatory within the scheme from 1 October 2026; before that the minimum was 2,400 dollars. From the same date 40% of the premium is to be paid to members in cash. (source 9)
Fairtrade also publishes a reference price it calculates a living income would need: 1,758 CFA francs a kilogram at the farm gate in Côte d'Ivoire and 45.4 cedis in Ghana, in April 2026. It is a reference figure, not a price anyone is obliged to pay. (source 10)
The two countries' systems smooth the price within a season and do not shield farmers from world prices between seasons, the living-income differential included, according to an interview study of the trade. (Staritz et al. 2022)
On the Living Income Differential, a model found income gains anywhere from zero to sizeable depending on how markets responded, and a later study of Côte d'Ivoire found no evidence that it had benefited farmers, nor any inefficiency in the chain that would explain low farm-gate prices. (Boysen et al. 2022; Kaestner et al. 2026)
What reached farmers after deductions, late payment or informal sale, season by season: no measurement was found.

Not shown: That any announced price was paid in full. An announcement is a promise.

Not shown: A comparison between the announced prices and the living-income reference prices. They are in different currencies at different dates under different definitions, and converting them here would manufacture a figure.

Not shown: What the differential did in Ghana. The one measurement read is for Côte d'Ivoire.

Certification

The subject with the widest gap between what is claimed and what could be verified. The best-known systematic reviews are not in the index used and were not read.

Where it is: certified farms worldwide are concentrated in areas that matter for biodiversity and not in the areas of deepest poverty. (Tayleur et al. 2018)
Prices: across seven tropical commodities, more certification went with larger trade volumes and, in most cases, lower export prices. Only standards with strict requirements and strong compliance checks produced a premium. (Bemelmans et al. 2026)
Forests: across 67 countries only one scheme coincided with lower loss of primary forest, three coincided with higher loss, and a finer spatial analysis found no effect of the standards. (Dröge et al. 2024)
Group membership, which certification usually requires, went with higher estimated yields in one Ghanaian comparison of 217 members and 199 non-members. (Donkor et al. 2023)
Across crops, a systematic review of 43 quantitative and 136 qualitative studies found certified producers got higher prices and more income from their produce, with no gain in overall household income and none in workers' wages, on evidence it judged thin. (Oya et al. 2017)
Cocoa is among the least studied certified crops, and Ivory Coast among the least studied countries. (Traldi 2021)
In Ghana and Ivory Coast, comparisons of UTZ-certified with other farmers at two dates found outcomes mixed and generally modest. Farmers given a full package of services gained in productivity and income; services had since fallen for half of them and the gains were levelling off. (Ingram et al. 2018)
In Côte d'Ivoire, Fairtrade raised household spending by 9%, and by 14% in poor households, on housing, clothing and the like. It had no significant effect on food consumption. In Ghana certified farmers were barely more food secure, and 65% of them remained vulnerable. (Knößlsdorfer et al. 2021; Dompreh et al. 2021)
In Indonesia the same label gave different results under different companies: income and production rose only in the schemes with the strongest interventions. (Bemelmans and Maertens 2025)
A randomised evaluation of cocoa certification: none was found. Every comparison above adjusts statistically for the fact that certified farmers differ from others to begin with.

Not shown: That certification fails, or that it works. The effects found are real and small, and they follow the services delivered more than the certificate.

Not shown: That a certified bar means a farmer above the poverty line. In the largest analysis read, most cocoa households were below a living income whatever their status.

Not shown: Anything specific to cocoa on price. The price study pools seven commodities.

Deforestation and traceability

Measured from satellites: cocoa is an underlying driver of over 37% of forest loss in protected areas in Côte d'Ivoire and over 13% in Ghana, and official figures understate the planted area, by up to 40% in Ghana. (Kalischek et al. 2023)
A second mapping found 20% of detected cocoa inside protected areas across the two countries, and cocoa in almost 70% of the protected areas studied. Its map was right about four times in five for land that is cocoa, and less often for land it labelled cocoa. (Abu et al. 2021)
Traders in several commodities, cocoa from Côte d'Ivoire among them, often buy more than 40% of their supply indirectly, outside the monitoring their sustainability claims rest on. (zu Ermgassen et al. 2022)
The European rule: cocoa is one of seven commodities covered. Products may be sold only if produced on land not deforested after 31 December 2020, legally produced, and covered by a due diligence statement, with the coordinates of every plot collected. (source 11)
Its start has moved twice: from 30 December 2024 to 30 December 2025, and then, under the latest amendment retrieved, to 30 December 2026, with 30 June 2027 for micro and small businesses. (source 11; source 12; source 13)
On whether better yields spare forest, the evidence points both ways. A scenario model estimated that earlier intensification could have avoided over 21,000 square kilometres of forest loss in West Africa; a survey of 1,035 Congo Basin households associated cocoa-based livelihoods with six to seven times more clearing. (Gockowski and Sonwa 2011; Ngouhouo-Poufoun et al. 2024)

Not shown: The share of cocoa that can be traced to a farm. The best-known estimate is in a paper available here only as a preprint, and was left out.

Not shown: What the European rule will cost farmers or traders, or whether it will reduce clearing. It had not applied when this was written.

Not shown: Anything later than the consolidated text of 18 September 2026, which still gives the dates above. The Commission's review of May 2026 proposed no change to them; a July 2026 act rewrote the product list and left cocoa as it was.

Where exported cocoa comes from: what can be traced, and who measures it

The question the deforestation rules turn on is whether a consignment can be tied to a place. One research initiative, Trase, publishes the most detailed public estimates for Côte d'Ivoire. Its figures are models built on disclosures, and are described here, not republished.

For 2019, fewer than half of Côte d'Ivoire's cocoa exports could be traced to a cooperative and a department: 43.6%. A further 23.9% was bought indirectly through local intermediaries, and 32.4% was exported by traders who disclosed nothing about their suppliers. (Renier, Vandromme and Meyfroidt 2023)
The same study attributes 2.4 million hectares of deforestation and forest degradation between 2000 and 2019 to cocoa, 45% of the country's total, and found that companies in the industry's forest initiative had mapped 40% of the farms supplying them, covering 22% of exports. (Renier, Vandromme and Meyfroidt 2023)
Trase now publishes that supply-chain dataset for 2016 to 2024, updated in May 2026, linking exports to departments of origin and to trading companies. (source 14)
Its numbers are estimates and say so. Production in each department is the national total shared out by mapped cocoa area and suitability. Deforestation is assigned to a year's exports by rule: a four-year lag from planting to harvest, and forest converted over fifteen years. (source 14)
Whether an exporter is counted as committed to zero deforestation is read from the company's own published wording. It is a record of what was promised, the first column of this page, not of what happened. (source 14)
The figures themselves are not reproduced on this site. Trase licenses each dataset separately, asks commercial users to make contact first, and says part of its underlying data was bought under restrictions on sharing. (source 15)

Not shown: Any department's or trader's figure. They are on Trase's own site, with its method notes; none was copied.

Not shown: How the picture has changed since 2019. The later years are in the dataset, which was not downloaded, and no published analysis of them was read.

Not shown: Comparable sub-national estimates for Ghana, Ecuador, Peru, Colombia, Indonesia or Brazil. Trase lists supply-chain data for each, older and without a published study read here.

Not shown: That a traced consignment is free of deforestation, or an untraced one linked to it. Traceability describes what is known about origin, not what happened there.

Farm economics: yield, efficiency and risk

Yields vary severalfold between farms in one country. Along Ghana's climate gradient they averaged 288 kg per hectare in the dry zone against 712 and 849 in the wetter zones, and across five Ivorian districts from 372 to 812. (Abdulai et al. 2018; Yéo et al. 2026)
What goes with higher yield: among 158 Ivorian farms, mineral fertiliser (29% higher), planting in rows (51%) and planting material of known origin (23%). These are associations in one survey. (Yéo et al. 2026)
Dependence on cocoa is highest where cocoa does best. Farmers in Ghana's wetter zones drew over 80% of their income from it; those in the dry zone had diversified. (Abdulai et al. 2018)
Indonesian farms ran at about half their attainable efficiency, though productivity had risen 75% in twelve years, mostly from more chemical inputs under subsidy. (Tothmihaly and Ingram 2019)
Women managing plots in Ghana were less efficient on average, a gap the authors link partly to fewer resources. (Danso-Abbeam et al. 2020)
Crop insurance was associated with higher income in a matched comparison of 600 Ghanaian farmers. (Agbenyo et al. 2022)
Price swings: over fifty years world price volatility had no significant link to cocoa supply in Ghana and a positive one in Nigeria. (Adegunsoye et al. 2024)
Diversifying pays those who have the land for it: among 303 Ivorian households, the highest incomes went with a second cash crop on a separate plot, and only 7% of producers had access to finance to invest in one. (Tokou et al. 2025)
Land tenure matters less simply than is assumed. Across 796 Ghanaian plots, farmers planted trees and improved soil even where they felt their rights insecure. (Asaaga et al. 2020)
Much cocoa is grown by people who do not own the farm. Hired labour and abusa, in which a caretaker works a landowner's farm, are long established in Ghana, with sharecropping traced to a shift after the 1920s, and the workers have little bargaining power. (Kissi and Herzig 2023; Gore 2025)
Replanting an old farm has no general rule: a review calls the decision complex, expensive and often contested between farmers and advisers. (Somarriba et al. 2021)
Where gold is mined illegally beside cocoa, 90 farmers in one Ghanaian district reported labour shortages, polluted water, competition for land and flooding from open pits. (Osman et al. 2022)
Production costs per hectare, days of labour, and returns to fertiliser measured on farms: still nothing with a verifiable abstract.

Not shown: That fertiliser, rows or better planting material cause the higher yields. Farmers who adopt them differ in other ways.

Not shown: Profit. Almost every study here measures yield or efficiency, not what the farmer has left.

Not shown: How many farms are worked by sharecroppers or caretakers, or what they earn. The studies read are qualitative.

Sources

Studies are cited by author and year and summarised, with their limits, on the research pages. The official documents are listed here. Each was downloaded and read, and the passage relied on was checked to be in it.

International Labour Organization. What is child labour (ILO topic page). www.ilo.org/topics/child-labour/what-child-labour
Official guidance, not law · International
NORC at the University of Chicago. NORC Final Report: Assessing Progress in Reducing Child Labor in Cocoa Production in Cocoa Growing Areas of Côte d'Ivoire and Ghana (2018/19 survey round) (October 2020). www.norc.org/content/dam/norc-org/pdfs/NORC%202020%20Cocoa%20Report_English.pdf
Official statistics · Côte d'Ivoire and Ghana
Living Income Community of Practice. Living Income Report, Rural Ghana: Cocoa growing areas of Ashanti, Central, Eastern, and Western Regions (Series 1, Report 1) (21 September 2018). www.cocoainitiative.org/sites/default/files/resources/LIVING-INCOME-REPORT-FOR-GHANA.pdf
Voluntary international text · Ghana
Living Income Community of Practice. Living Income Benchmark, June 2022 Update, Côte d'Ivoire, Rural cocoa growing areas (June 2022). www.living-income.com/fileadmin/user_upload/2022_Update_Living_Income_Report_Cote_d_Ivoire_Cocoa.pdf
Voluntary international text · Côte d'Ivoire
KIT Royal Tropical Institute. Demystifying the cocoa sector in Ghana and Côte d'Ivoire (2018). www.kit.nl/wp-content/uploads/2020/05/Demystifying-complete-file.pdf
Official statistics · Ghana and Côte d'Ivoire
KIT Royal Tropical Institute. Analysis of the income gap of cocoa producing households in Ghana: Comparison of actual incomes with the Living Income Benchmark (11-09-2018). www.cocoainitiative.org/sites/default/files/resources/INCOME-GAP-FOR-COCOA-GROWING-HH-IN-GHANA.pdf
Official statistics · Ghana
Ghana Cocoa Board (COCOBOD). Review of the Producer Price of Cocoa for the 2026/27 Cocoa Season (25th September 2026). cocobod.gh/news/review-of-the-producer-price-of-cocoa-for-the-202627-cocoa-season
Law · Ghana
Conseil du Café-Cacao. Conseil du Café-Cacao website, home page price panel (07 Octobre 2026). www.conseilcafecacao.ci/
Law · Côte d'Ivoire
Fairtrade International. Cocoa Price announcement from Fairtrade International - December 2025 (11.12.2025). www.fairtrade.net/content/dam/fairtrade/fairtrade-international/standards/standards-and-pricing-work-in-progress/cocoa/2025-12-11_Cocoa-Price-Announcement_EN.pdf
Voluntary international text · International
Fairtrade International. Fairtrade Living Income Reference Prices for Cocoa from Ghana & Côte d'Ivoire: Explanatory Note (April 2026). www.fairtrade.net/content/dam/fairtrade/fairtrade-international/library/2026/west-africa-cocoa-living-income-reference-prices-2026/Living-income-reference-price-cocoa-ghana-cote-divoire-explanatory-note-2026.pdf
Voluntary international text · Côte d'Ivoire
European Parliament and Council of the European Union. Regulation (EU) 2023/1115 of the European Parliament and of the Council on the making available on the Union market and the export from the Union of certain commodities and products associated with deforestation and forest degradation (31 May 2023). eur-lex.europa.eu/eli/reg/2023/1115/oj
Law · EU
European Parliament and Council of the European Union. Regulation (EU) 2024/3234 amending Regulation (EU) 2023/1115 as regards provisions relating to the date of application (19 December 2024). eur-lex.europa.eu/eli/reg/2024/3234/oj
Law · EU
European Parliament and Council of the European Union. Regulation (EU) 2025/2650 amending Regulation (EU) 2023/1115 as regards certain obligations of operators and traders (19 December 2025). eur-lex.europa.eu/eli/reg/2025/2650/oj
Law · EU
Trase (Stockholm Environment Institute and Global Canopy). Trase open data: cocoa datasets listing (2026-05-11). trase.earth/open-data
The organisation's own account · International
Trase (Stockholm Environment Institute and Global Canopy). Trase: Terms of Use (2025-05-04). trase.earth/terms-of-use
The organisation's own account · International

Reviewed 9 October 2026.