Social Protection

Bangladesh Social Protection and Safety Nets

Executive Summary

Bangladesh operates one of the largest social safety net systems in the developing world, with 113 registered programs, a budget of BDT 1130 billion (2.53% of GDP), and coverage reaching an estimated 30.5 million households. Yet this vast apparatus suffers from three structural failures that severely limit its poverty reduction impact: extreme fragmentation across 23+ ministries with no effective coordination mechanism, targeting errors that leave 71.0% of the poor without any benefit while 46.7% of recipients are non-poor, and benefit levels so low (BDT 500/month for old age allowance, roughly $4.55) that they cannot meaningfully reduce poverty even when they reach the right people. The National Social Security Strategy (NSSS), approved in 2015, laid out an ambitious lifecycle framework for reform, but implementation has been slow, selective, and inadequately funded. The gap between design and delivery defines Bangladesh's social protection challenge.

Scale and Architecture

Bangladesh's social protection system is vast in scope but fragmented in execution. The 113 registered safety net programs span cash transfers, food distribution, employment generation, education stipends, microcredit, and disaster relief. They are administered by at least 23 ministries and divisions, with the Department of Social Services (DSS), the Ministry of Disaster Management, the Ministry of Food, the Ministry of Primary and Mass Education, and the Ministry of Women and Children's Affairs each running major programs with minimal coordination. The total budget of BDT 1130 billion (2.53% of GDP) is substantial in absolute terms but modest relative to the scale of poverty and vulnerability. By comparison, India allocates approximately 1.5% of GDP to direct safety net transfers, Nepal approximately 2.8%, and Sri Lanka approximately 1.2%.

The fragmentation creates several pathologies. Multiple programs serve overlapping populations while leaving entire categories of vulnerable people uncovered. Administrative costs are multiplied across parallel delivery systems. No single ministry has an authoritative picture of who receives what, making deduplication impossible without a functioning single registry. The NSSS envisioned consolidating programs into five lifecycle clusters: children, working-age, elderly, disabled, and shock-affected. A decade after approval, the core institutional architecture for this consolidation, particularly the single registry and unified MIS, remains incomplete.

Coverage and Benefit Adequacy

The old age allowance (Boysoko Bhata) is Bangladesh's flagship social pension, reaching 5.7 million beneficiaries, approximately 34.0% of the population aged 60 and above. At BDT 500 per month (roughly $4.55), the benefit is critically inadequate by any measure. The amount has not been meaningfully increased since 2016, losing approximately 35-40% of its real value to cumulative inflation over that period. Nepal's universal old age pension, by contrast, provides NPR 4,000/month (roughly $30) and covers all citizens aged 68+, while India's Indira Gandhi National Old Age Pension Scheme provides INR 200-500/month but only to below-poverty-line households.

The widow allowance reaches 2.46 million women, and the disability allowance covers 2.05 million beneficiaries, both at the same BDT 500-800/month range. Primary education stipends reach 13.0 million students and secondary stipends cover an additional 4.2 million, making the stipend programs the largest single category by beneficiary count. The stipends have been credited with driving Bangladesh's remarkable gains in female enrollment, particularly at the secondary level, where female enrollment now exceeds male enrollment in a reversal of the historical pattern.

The fundamental problem across all programs is that benefit levels were set low to maximize coverage breadth, a politically rational but analytically flawed approach. A transfer of BDT 500/month represents approximately 3-4% of the poverty line income, far below the threshold at which transfers begin to produce measurable impacts on consumption, nutrition, or human capital investment. The result is a system that distributes small amounts to many people rather than adequate amounts to those who need it most.

Targeting Efficiency

Targeting is the central failure of Bangladesh's safety net system. The exclusion error of 71.0% means that the majority of poor households receive no safety net benefit whatsoever, while the inclusion error of 46.7% means that nearly half of all recipients are non-poor. The overall leakage rate is estimated at 24.0%. Targeting is severely compromised.

The root causes are structural, not merely administrative. Beneficiary selection at the union parishad level is dominated by local political considerations: elected officials allocate slots to supporters, relatives, and influential community members rather than strictly by poverty criteria. Proxy means testing (PMT), which uses observable household characteristics to predict income, performs poorly in the Bangladesh context because the variance in consumption among households near the poverty line is small, making accurate classification statistically difficult. Community-based targeting, which relies on local knowledge of poverty, is compromised by the same political economy that distorts formal selection processes.

The case for moving toward categorical or universal approaches is strengthening. Universal old age pension (covering all citizens above a defined age), universal child benefit, and universal disability allowance would eliminate exclusion errors by definition, remove the political economy of selection, and reduce administrative costs associated with targeting. The fiscal cost of universality is higher in gross terms but lower than commonly assumed when targeting costs and leakage are accounted for. Nepal's universal old age pension demonstrates that universality is feasible at Bangladesh's income level.

Food-based Safety Nets and Employment Programs

Food-based programs remain a major component of Bangladesh's safety net architecture, reflecting both the country's history of famine and food insecurity and the political salience of food distribution. The Vulnerable Group Development (VGD) program reaches 1.05 million women with food rations plus development training. The Vulnerable Group Feeding (VGF) program provides food transfers to 5.0 million households during lean seasons and after disasters. The Food Friendly Program (OMS/TCB) sells rice at subsidized prices to 5.0 million card-holders.

The Employment Generation Programme for the Poorest (EGPP) provides 0.8 million workers per cycle with 80 days of employment in rural infrastructure maintenance at BDT 200/day. EGPP is Bangladesh's closest equivalent to India's MGNREGA (Mahatma Gandhi National Rural Employment Guarantee Act) but operates at a fraction of the scale: MGNREGA provides a legal guarantee of 100 days employment to any rural household that demands it, while EGPP rations limited slots through the same targeting mechanisms that produce high exclusion errors. Scaling EGPP toward an employment guarantee would provide a powerful counter-cyclical and climate-adaptive safety net, as demand for work rises precisely when agricultural income falls.

G2P Digitization and MIS Integration

Bangladesh is making meaningful progress on digitization, with 42.0% of safety net payments delivered through mobile financial services (primarily bKash and Nagad) or bank transfers. This shift from manual cash distribution has reduced several forms of leakage: intermediaries who historically skimmed payments, ghost beneficiaries who existed only on paper, and delays that eroded the real value of transfers. The COVID-19 pandemic accelerated digitization, as emergency cash transfers to 5 million households were delivered via mobile money within weeks.

However, significant challenges remain. The National Household Database (NHD), intended to serve as a single registry for all safety net programs, covers approximately 35 million households but suffers from data quality problems: outdated records, incomplete coverage of urban areas, and limited capacity for dynamic updating as households move in and out of poverty. Without a functioning single registry, deduplication across programs is impossible, and the goal of a consolidated lifecycle transfer system remains aspirational. The MIS systems of individual programs (DSS, MoPME, Ministry of Food) operate on separate platforms with limited interoperability.

Financial inclusion linkages are underexploited. The 180+ million mobile money accounts in Bangladesh represent a delivery infrastructure of extraordinary reach, but most safety net beneficiaries use mobile wallets only for cash-out, not for savings, insurance, or credit access. Linking G2P payments to financial products, micro-savings with matched incentives, micro-insurance for health and crop loss, credit scoring based on payment history, could transform the safety net from a passive transfer system into an active financial inclusion platform.

NSSS Implementation and Reform Agenda

The NSSS, approved in 2015 with World Bank and DFID support, remains the most ambitious social protection reform framework in South Asia. Its lifecycle approach envisions five pillars: (1) a child benefit from birth to age 4, (2) school stipends and feeding from age 5-18, (3) working-age employment, insurance, and skills programs, (4) old age pension, and (5) disability and vulnerability-based support. The strategy also called for a single registry, consolidated MIS, strengthened monitoring, and a shift from project-based to program-based budgeting.

A decade into implementation, progress is uneven. The universal pension scheme (Probash), launched in August 2023, is the most significant structural reform: it offers contributory pension products to all Bangladeshi citizens aged 18-50, with government co-contribution for informal sector workers earning below a defined threshold. Early enrollment of 0.3 million is modest but represents a genuine institutional innovation. The critical design question is whether a voluntary, contributory scheme can achieve meaningful coverage among informal workers (who constitute 85%+ of the labor force) without stronger incentives or mandates.

The most conspicuous gap in the lifecycle framework is the absence of a universal child benefit. The evidence base for early childhood nutrition and stimulation investment is overwhelming: returns to investment in the first 1,000 days exceed returns at any subsequent life stage. Bangladesh spends approximately 8% of its social protection budget on children despite children constituting 30%+ of the population. A universal child benefit of BDT 800/month for children under 5 would cost approximately 0.4% of GDP, a feasible fiscal commitment with transformative potential for stunting reduction, school readiness, and intergenerational poverty reduction.

BRAC's Ultra-Poor Graduation model, developed in Bangladesh and now replicated in over 40 countries, provides the most rigorous evidence base for sustainable poverty exits. The model combines asset transfer, skills training, consumption support, savings facilitation, and coaching over a 24-month period. Randomized evaluations show lasting income gains 7+ years after program completion. The government's adaptation of this model through the "My House, My Farm" program has reached millions of households but with diluted intensity and quality control compared to the original BRAC design.

Shock-Responsive Social Protection

Bangladesh's extreme climate vulnerability, annual flooding affects 20-30% of the country, cyclones strike the coastal belt every 2-3 years, riverbank erosion displaces 100,000+ people annually, makes shock-responsive social protection not a luxury but a necessity. The concept of adaptive social protection involves two mechanisms: vertical expansion (increasing benefit amounts to existing beneficiaries during shocks) and horizontal expansion (temporarily enrolling new beneficiaries affected by a specific shock).

Current practice is ad hoc: after cyclones or floods, the government distributes VGF allocations and GR (Gratuitous Relief) through local disaster management committees. These responses are slow, politically driven in allocation, and disconnected from the regular safety net system. A pre-positioned, rules-based system, where automatic triggers (flood levels, cyclone categories, food price thresholds) activate predetermined expansions of existing programs, would be faster, more transparent, and more equitable. Forecast-based financing, which disburses funds based on weather forecasts rather than post-disaster damage assessments, has been piloted by the World Food Programme in Bangladesh with promising results.

Outlook, Risks, and Policy Recommendations

Three risks define the social protection outlook:

• Fiscal space compression: Bangladesh's tax-to-GDP ratio of approximately 8%, among the lowest globally, severely constrains the fiscal envelope for social protection. Without revenue mobilization, the 2.5% of GDP allocation cannot grow to match the scale of need, and real benefit values will continue to erode through inflation.
• Demographic transition pressure: Bangladesh's population is aging rapidly, with the 60+ cohort projected to reach 22% by 2050. The current old age allowance architecture, low coverage and negligible benefits, is wholly inadequate for an aging society. Without a functioning pension system (either contributory or tax-financed), old age poverty will emerge as a major social crisis within two decades.
• Climate-driven poverty reversals: Climate shocks are pushing vulnerable households back below the poverty line faster than safety nets can protect them. The poverty rate of 18.7% and extreme poverty rate of 5.6% mask significant vulnerability: an estimated 15-20% of the population lives within 10% of the poverty line and can be pushed below it by a single flood, illness, or income shock.

Five policy recommendations:

• Introduce a universal child benefit of BDT 800/month for children under 5, financed through reallocation from less effective programs and modest fiscal expansion. Estimated cost: 0.4% of GDP. Expected impact: reduced stunting, improved school readiness, and lower intergenerational poverty transmission.
• Index all social protection benefits to inflation, with automatic annual adjustment based on the consumer price index. The erosion of the BDT 500 old age allowance by 35-40% in real terms since 2016 is a policy failure that can be prevented through indexation.
• Complete the National Household Database and single registry, with mandatory interoperability across all program MIS systems. This is the prerequisite for deduplication, improved targeting, and lifecycle coverage tracking. Allocate dedicated budget and institutional authority to the Cabinet Division's Central Management Committee.
• Scale the EGPP toward a rural employment guarantee, with legally mandated minimum days of employment per household per year. Link to climate adaptation by directing employment toward flood embankment maintenance, drainage improvement, and community infrastructure resilience.
• Establish a shock-responsive social protection framework with pre-agreed triggers, financing mechanisms, and horizontal/vertical expansion protocols linked to the disaster management system. Pre-position 0.2% of GDP as a contingency fund for automatic disbursement when triggers are activated.

*Data sources: Ministry of Finance Budget Documents FY2023-24, World Bank Bangladesh Social Protection Public Expenditure Review 2023, HIES 2022, DSS Annual Report 2023, NSSS Mid-Term Review 2021, BRAC Research.*