Spatial Poverty in Bangladesh: Evidence from HIES, IPUMS, and DHS

Executive Summary

Bangladesh's national poverty headcount ratio is 18.7% (HIES 2022, upper poverty line), with extreme poverty at 5.6% and a Gini coefficient of 0.499. National figures mask a severe spatial divide. Mymensingh records the highest divisional headcount at 28.7%; Dhaka records the lowest at 10.8%. The 17.9 percentage-point gap is substantial and cannot be closed by a nationally uniform policy instrument. Special geographic areas carry poverty rates two to three times the national average: Chittagong Hill Tracts (65.0%), chars (52.0%), haors (45.0%), and the coastal belt (35.0%). Each is a distinct poverty trap requiring a tailored institutional response. With GNI per capita at $2,824 and a population of 174 million, Bangladesh's residual poverty challenge is geographic, not aggregate.

Bottom line: Uniform national programs are the wrong instrument. Geographic targeting, special area development authorities, monga mitigation, and CHT land reform are the four highest-return interventions available.

National Poverty Profile

The national headcount ratio fell from 48.9% in 2000 to 18.7% in 2022 (BBS HIES). Rural poverty declined from 52.3% to 20.5%; urban poverty from 35.2% to 14.7%.

The rural-urban divide persists: rural poverty at 20.5% exceeds urban at 14.7%, but urban measurement understates deprivation among the 5.0 million slum residents in Dhaka and Chittagong. Conventional income poverty lines do not capture the housing, sanitation, and tenure insecurity those households face.

The Gini of 0.499 signals that aggregate poverty reduction has not produced shared prosperity. Growth benefits have concentrated in the Dhaka-Chittagong corridor; lagging regions have received aggregate trickle-down without structural transformation. The multidimensional poverty index (national MPI: 24.1%) confirms that the remaining poverty is multi-dimensional: income alone understates deprivation in asset-poor, education-deficient, and employment-scarce districts.

Division-Level Disparities

HIES 2022 division headcount ratios (upper poverty line): Mymensingh: 28.7%; Rangpur: 24.8%; Barisal: 22.4%; Rajshahi: 20.4%; Khulna: 17.5%; Sylhet: 17.4%; Chittagong: 13.1%; Dhaka: 10.8%.

Average monthly household income in Dhaka (BDT 42,156) is 1.9x that of Rangpur (BDT 21,890), a structural income gap that survey-based poverty headcounts understate.

The high-poverty divisions share a structural profile: geographic distance from the Dhaka-Chittagong economic corridor, a single-crop rice agricultural calendar, and deep exposure to the annual monga (September-November lean season) that affects 6 districts in the northwest. During monga, agricultural wage labor disappears before the boro harvest; households deplete assets, reduce caloric intake, and send members to urban areas on distress migration. The economic cost is permanent: monga exposure in childhood is associated with measurable losses in adult productivity.

The Sylhet case illustrates that remittances are not a structural solution. Despite the highest per-capita diaspora flows in Bangladesh, Sylhet records poverty and child stunting rates near or above the national average. Remittances transfer income to specific households; the haor ecology (4-6 months annual submergence) prevents the infrastructure investment and agricultural diversification that would generate broad-based local income. Transfer income without structural transformation produces consumption without development.

Special Area Poverty: Four Distinct Traps

Each of the following zones has a poverty rate more than double the national average and requires a dedicated institutional response. Bundling them into a generic lagging-region program is the primary historical policy failure.

Chittagong Hill Tracts (65.0%): Topographic isolation, ethnic marginalization of indigenous communities (Chakma, Marma, Tripura, and others), and the 1997 Peace Accord's incomplete implementation have produced a poverty trap that persists across development cycles. The CHT Land Commission has resolved a small fraction of submitted land disputes in over two decades. Land insecurity is the binding constraint; no social protection program resolves it.

Char Areas (52.0%): Riverine island communities occupy ephemeral land created and destroyed by erosion and accretion. Tenure insecurity, absent infrastructure, and immature soils suppress investment. The Char Development and Settlement Project (CDSP) has demonstrated that surveyed land allocation, raised homesteads, and agricultural extension can reduce char poverty materially, but coverage remains a fraction of the char population.

Haor Areas (45.0%): Seasonal submergence for 4-6 months annually compresses agriculture into a single boro rice crop and makes standard infrastructure economically marginal. The Haor Master Plan (2012) correctly diagnosed the constraint; implementation has been fragmented and underfunded.

Coastal Belt (35.0%): Salinity intrusion (reducing crop yields and drinking water access), cyclone exposure, and sea-level rise compound income poverty. Climate projections place this zone at the front line of displacement risk by 2050. Poverty here is already partly climate poverty; adaptation and protection must be built into the intervention design.

District-Level Deprivation: IPUMS MPI

District-level MPI data is unavailable from the current data source (IPUMS 2011 CSV absent on this deployment).

The MPI decomposition across tracked districts (coverage: data unavailable) shows asset deprivation as the dominant dimension, followed by education. Bangladesh has achieved near-universal primary enrollment, but the quality of education and returns in asset-poor, geographically isolated districts remain insufficient to break intergenerational poverty transmission. Employment deprivation, while lower in aggregate, is severely concentrated in monga-affected and char/haor districts where seasonal labor markets collapse for months.

Sylhet records the highest child stunting rate at 33.9% (DHS 2022), with wasting at 12.2%. These figures confirm that poverty and child malnutrition are spatially co-located and mutually reinforcing.

Satellite Cross-Validation

Satellite-derived nightlights (mean radiance: 0.9), built-up area (2,678 km2), and vegetation index (NDVI: 0.530) independently corroborate the survey-based spatial patterns: low nightlight intensity is tightly correlated with high divisional poverty rates, confirming that economic activity, electrification, and deprivation are spatially co-determined. The World Bank poverty headcount (18.7%) and GNI per capita ($2,824) are consistent with HIES estimates, confirming Bangladesh's lower-middle-income status with above-average spatial dispersion of deprivation.

Recommendations

The following four interventions are prioritized by tractability, impact, and grounding in the evidence above. They are not generic aspirations; each has a specific institutional mechanism and a measurable target.

1. Geographically differentiated social protection allocation. Social protection coverage at 28.5% is both insufficient and poorly targeted spatially. The 17.9 percentage-point divisional headcount range demands that the allocation formula for allowances, public works, and school feeding assign weights proportional to divisional poverty rates. BBS HIES 2022 division data should be the direct input to this formula, updated with each HIES cycle.

2. Four special-area development authorities with statutory mandates. CHT, chars, haors, and the coastal belt each need a dedicated authority with a multi-year budget, a technical mandate covering land, infrastructure, and livelihoods, and performance accountability to a measurable poverty target. The current project-based fragmentation (multiple ministries, no single accountable entity) is the structural cause of repeated implementation failure. The CDSP model and the Haor Master Plan provide institutional templates.

3. Northwest monga mitigation and agricultural diversification. The 6 monga-affected districts need: (a) a guaranteed public works employment program covering the September-November lean season at a wage that prevents asset depletion; (b) agricultural extension for crop diversification beyond boro rice; and (c) cold storage and agro-processing investment that creates year-round non-farm employment. Each element has precedent (MGNREGA in India, CDSP agricultural extension in chars); none requires institutional innovation, only prioritization and funding.

4. CHT land dispute resolution as a precondition for poverty reduction. No economic intervention in the CHT will achieve durable results while land tenure remains contested. The government should set a binding timeline for the CHT Land Commission to resolve the backlog of disputes, protect indigenous land from encroachment under a statutory framework, and fund infrastructure and services sized for hillside communities. This is a political commitment as much as a budget decision.

Sources: BBS HIES 2022; IPUMS Bangladesh Census 2011; DHS 2022; OPHI National MPI 2025; NASA VIIRS; GHSL 2023; MODIS NDVI; World Bank WDI; CDSP; Haor Master Plan 2012; ICZM.