Cocoa: two readings that do not meet

The household survey and the industry’s monitoring systems give very different readings of child labour in West African cocoa.

Child labour in West African cocoa: two readings, each on its own base

45%Household survey (NORC), 2018/19 seasonChildren aged 5–17 in agricultural households in the cocoa-growing areas of Côte d’Ivoire and Ghana
11%Industry monitoring systems (ICI members), season to 30 September 2025Children covered by ICI members’ monitoring and remediation systems in Cameroon, Côte d’Ivoire, Ghana and Nigeria

Six years apart, different countries and populations: no ratio of one to the other is drawn.

Children found in child labour by ICI-implemented systems who were no longer in it at their two most recent follow-up visits% of children receiving a second follow-up visit that year

The programme’s own reporting.

The two readings

Reading	Child labour, %	Base	Where	Period
NORC at the University of Chicago (funded by US Department of Labor)	45	% of children 5-17 in agricultural households in cocoa-growing areas	Côte d'Ivoire and Ghana, cocoa-growing areas (combined)	2018/19 season
International Cocoa Initiative (ICI)	11	% of children covered	Cameroon, Côte d'Ivoire, Ghana and Nigeria (ICI members)	cocoa season to 30 September 2025

The follow-up series

Year	No longer in child labour at follow-up, %
2018	27.4
2019	35.5
2020	35.8
2021	41
2022	47.4
2023	47.3

What this joins

NORC at the University of Chicago, Assessing Progress in Reducing Child Labor in Cocoa Production in Cocoa Growing Areas of Côte d'Ivoire and Ghana, Final Report (October 2020). Table 6 (printed p.49): Prevalence Rates; Estimate (95% CI)
International Cocoa Initiative, ICI Annual Report 2025; accessed 15 October 2026. PDF p.5 (printed p.8)
International Cocoa Initiative, ICI Annual Report 2023 (published June 2024). PDF p.15 (printed p.28), bar chart '% no longer in child labour at two most recent follow-up visits' (bar labels matched to year axis by position)

How it was computed

Each reading is shown as its publisher printed it, with its own base, population and period.

No ratio of one reading to the other is computed.

The follow-up series is read from the bar chart in ICI’s 2023 annual report, each bar matched to its year by position.

What it does not show

Six years apart, different countries and populations: no share of the survey’s estimate reached by the systems is computed.
The 11% and the follow-up series are the programme’s own reporting (limited assurance by PwC on the reporting process, not the outcomes).

Read the method note Back to Analysis