Bangladesh's Multiple Indicator Cluster Survey 2025 (MICS 2025), released by the Bangladesh Bureau of Statistics (BBS) and UNICEF on 16 November 2025, recorded child labour at 9.2 percent of children aged 5 to 17 -- up from 6.8 percent in MICS 2019. The survey, which covered nearly 63,000 households across all divisions, districts, and three city corporations, is the most geographically granular child welfare census the country has produced.

The arithmetic is unambiguous. A 35.3 percent rate increase translates to approximately 1.2 million additional children in the labour force over six years. UNICEF described this as "a nearly 40 percent increase in child labour -- children out of school, trapped in poverty, and unable to escape the cycle." The MICS frame uses the ILO operational definition: for children aged 5 to 11, any economic activity; for ages 12 to 14, more than 14 hours per week or hazardous conditions; for ages 15 to 17, more than 43 hours per week or hazardous work. Household chores below those thresholds are excluded. The 2019 and 2025 rounds used the same definition, so the comparison is methodologically sound.

The reversal is stark in historical context. Between 2000 and 2019, Bangladesh cut its child labour rate by roughly half through a combination of targeted conditional cash transfers, school stipend programmes, and export-sector compliance pressure from international buyers. The post-2020 period reversed that trajectory: the COVID-19 shock, subsequent inflation, and declining real wages in low-income households pushed children back into earnings roles. The MICS 2025 does not publish a cause attribution, but the timing is consistent with observed poverty dynamics.

Where Children Work: The Informal Sector Problem

The Bangladesh National Child Labour Survey 2022 (NCLS 2022), conducted by BBS with ILO support, provides the most recent sector breakdown. By headcount: services absorb approximately 1.27 million child workers, industry and manufacturing approximately 1.19 million, and agriculture approximately 1.08 million (23.6 percent of the total). All three categories are dominated by informal, unregistered settings: domestic service, tea stalls, transport and auto-repair workshops, brick-breaking, shrimp processing, and seasonal agricultural labour.

This matters for enforcement. The Department of Inspection for Factories and Establishments (DIFE) -- MoLE's field inspection arm -- has 724 inspector positions across 31 district offices, a major expansion from the approximately 250 total officials the department had before the Rana Plaza collapse in 2013. But DIFE's mandate covers registered factories and establishments. Domestic workers, agricultural labourers, and children in informal workshops are structurally outside routine inspection. The MoLE's revised Hazardous Occupations List (43 sectors as of April 2022, expanded from 38 in 2013) defines prohibited conditions with precision. Without a mechanism to reach informal settings, the list is a prohibition without enforcement reach.

The Violence Overlay

MICS 2025 also recorded that 86 percent of children aged 1 to 14 experienced violent discipline -- physical punishment or psychological aggression -- in the month before the survey. This is a population-wide figure across all households, not a cross-tabulation with labour status. The Children's Act 2013 prohibits corporal punishment in educational institutions but does not extend the prohibition explicitly to domestic or informal workplace settings. Research in comparable South Asian contexts consistently finds that households which permit child labour also exhibit higher tolerance for physical discipline, and that working children face elevated violence exposure relative to non-working peers. The co-occurrence of economic exploitation and physical discipline in the same household environments is a compounding deprivation, not two separate problems.

The Formal-Informal Divergence

Bangladesh's best child labour compliance performance is concentrated in the most visible and internationally audited segment of the economy. BGMEA's zero-tolerance policy, adopted in 1995 alongside UNICEF and ILO, has produced near-zero child labour in direct-hire ready-made garment factories. International buyer compliance audits and the Accord/RSC monitoring framework enforce this position in the formal sector.

The January 2025 GoodWeave and Business and Human Rights Centre study punctured the headline claim. It found child labour present in RMG export supply chains via subcontracted and informal units that produce inputs for formal export factories. All minors interviewed were illegally employed. The RSC Complaints Mechanism expanded in November 2025 to cover broader labour violations, but its jurisdiction does not extend to unregistered subcontractors. The formal-informal gap in compliance mirrors the formal-informal gap in inspection: robust enforcement where enforcement is easiest, minimal reach where the problem is concentrated.

Four Structural Interventions

Extend inspection jurisdiction to informal workplaces. DIFE's expansion post-Rana Plaza was sector-specific and factory-focused. A parallel enforcement model for informal settings -- drawing on community-level social service workers, union-level upazila monitoring cells, and MoLE data-sharing with the Department of Social Services -- does not require new legislation. The Children's Act 2013 already provides the legal basis; the gap is operational reach.

Anchor the school stipend programme to child-labour-at-risk households. The primary mechanism linking child labour and school dropout is household income pressure at the margin. Bangladesh's Primary Education Stipend Programme (PESP) and Secondary School Stipend Programme (SSSP) already cover a large share of low-income children. Retargeting a meaningful fraction of stipend expansion toward households identified as child-labour-at-risk in the MICS 2025 divisional data would directly address the price signal that draws children out of school and into the labour market.

Mandate supply chain transparency to the subcontractor level in export industries. The formal RMG sector's compliance performance demonstrates what buyer pressure and audit coverage can achieve. Extending that model -- requiring export factories to register their subcontractors and accept audit coverage -- is a supply-chain governance measure, not a new regulatory burden on the formal sector. The RSC has the infrastructure; the scope needs to be widened by MoLE directive.

Use MICS 2025 divisional estimates to target intervention geography. The MICS 2025's district-level resolution is an operational asset that has not existed in previous survey rounds. Divisions and districts with the highest child labour rates should be the first recipients of additional DIFE inspectors, stipend expansion, and community-level social protection outreach. Geographic targeting converts a national average into an actionable programme map.