Can social protection budget deliver?

The FY2026-27 social security budget is the largest in Bangladesh's history. At Tk 1,44,338 crore ( or Tk 1.44 trillion), it represents a 14 per cent nominal increase, presented by Finance Minister Amir Khosru Mahmud Chowdhury in Parliament as evidence of decisive movement toward the government's commitment to raise social protection spending to 3 per cent of gross domestic product (GDP) by FY2027-28. The reported figure this year is 2.11 per cent of GDP.
The right question is not whether this budget is large. It is whether it is strategic and fiscally credible, whether it moves Bangladesh structurally closer to that 3 per cent target, or whether it is largely reclassification and a new flagship programme without a clear financing roadmap.
HOW MUCH OF THIS IS ACTUALLY SOCIAL SECURITY: Of the 90 programmes in this year's budget, the Finance Division's own classification identifies only 48 as genuinely "pro-poor": programmes whose primary design is poverty reduction through direct benefits to identified poor households. These 48 receive Tk 562.29 billion, just 38.9 per cent of the headline allocation. The largest item outside this block is Pension Management, with Tk 353.79 billion for 9.59 lakh retired civil servants, consuming roughly two thirds of what all 48 pro-poor programmes receive combined.
There is nothing inherently wrong with a civil service pension. It is a contractual obligation, no different in kind from a private firm's pension liability. The problem is presenting it as social security, which obscures the true scale of anti-poverty spending. When citizens hear Bangladesh spends Tk 1.44 trillion on social security, they reasonably imagine support reaching poor households. In reality, less than 40 per cent of that envelope does.
This changes the denominator. If genuine anti-poverty spending is roughly 39 per cent of a 2.11 per cent of GDP envelope, the share actually reaching poor households is likely in the 0.8 to 1 per cent range, far from the 3 per cent target set for the entire social security envelope, pensions included. Bangladesh does not only need a larger social protection budget; it needs a clearer one, so the public, investors, and development partners can distinguish poverty reduction spending from a legitimate but separate pension obligation.
The harder question is what closing the gap to 2 or 3 per cent would cost, and where the money comes from. As I have argued elsewhere, raising Bangladesh's tax to GDP ratio by around two percentage points would, in principle, generate additional revenue equivalent to roughly 2 per cent of GDP, enough to finance that expansion without deficit financing. Reallocation and borrowing are the other two channels, but each carries a cost. Reallocation means explicit, currently unstated trade-offs against other ministries, and sustained borrowing raises debt-servicing burdens for future budgets. None of the three paths is named in this year's budget.
A PORTFOLIO BEING REBALANCED WITHOUT AN EXPLANATION: Within the pro-poor block, this year represents a sharp rebalancing, not a uniform expansion. The Family Card Programme rises from a pilot allocation of Tk 866.10 million to Tk 145 billion in a single cycle. Financing that scale-up required reallocating shares within the pro-poor envelope: public workfare fell from 6.85 to 4.89 per cent of social assistance, education stipends from 10.64 to 8.06 per cent, and food distribution from 9.80 to 7.13 per cent. The Household-Shock category, which buffers families against floods and crop failures, was cut by about 7 per cent, and school-age children's funding fell by about 4 per cent.
The more useful way to read this is as a change in the portfolio's risk profile. Workfare and food distribution are countercyclical instruments that expand when shocks hit. Stipends and early childhood spending are human capital investments with long payoff horizons. Household cash transfers, by contrast, are recurring entitlements with a fixed monthly cost, largely independent of whether a shock has occurred. Shifting the centre of gravity from the first two toward the third means less automatic disaster buffering, less investment in future productivity, and more fixed recurring liability.
That trade-off may well be right. International evidence shows well designed cash transfers can reduce poverty efficiently while giving households flexibility. But the budget offers no visible analysis of why these particular programmes were chosen for contraction, or what happens to the households who depended on them.
THE MISSING MEDIUM-TERM PLAN: The budget document is unusually candid about the scale of the remaining task. Its review of a decade under the National Social Security Strategy (NSSS, 2015 to 2026) shows budget adequacy as the one dimension where progress has been negligible. The target was 3 per cent of GDP; the outturn is 2.11 percent. It also confirms that targeting remains weak, with only about half of poor households receiving assistance in 2022 while 22 percent of benefits went to the wealthiest fifth, and that contributory social insurance for the 85 percent employed informally remains, in the document's own words, "near total absence."
What is missing is the bridge between diagnosis and delivery. The current allocation covers 4.10 million households. But the government's own Family Card Piloting Implementation Guideline, 2026, envisions gradual expansion to 20 million households, evolving into a "Universal Social ID" by 2030. As I noted in earlier commentary on this programme, that scale of rollout would cost roughly Tk 600 billion a year, about 1 per cent of GDP on its own, even before health, education, and ageing-related pension pressures are factored in.
None of this costing appears in the budget itself, and there is no expenditure framework linking the 3 per cent target to specific revenue measures, nor any account of how much of the path will come from new revenue versus reclassification. The government's own NSSS-II framing (FY2026-27 to FY2035-36), covering consolidation, full PMT deployment, CPI-linked indexation, and contributory extension, is, in substance, the right agenda. What it lacks is a financing plan with dates attached.
BUILDING SYSTEMS, NOT JUST PROGRAMMES: The Family Card could become the foundation of a more integrated system built around a dynamic registry, digital payments, and transparent eligibility. That would be a meaningful shift from a system that has historically evolved programme by programme. But building systems requires governance, and two choices will determine whether it becomes durable infrastructure.
The first is registry governance. The Family Card combines PMT with ward-level verification committees, which is sensible, since PMT is more accurate using observable assets while local input catches households whose circumstances change suddenly. But the design is only as credible as the separation between who sets the rules and who disburses the money. Large social protection systems work best when targeting methodologies and registries are subject to independent oversight and regular scrutiny. Brazil and Indonesia have both refined their social registries through audits and institutional reform to improve transparency and reduce inclusion and exclusion errors.
Bangladesh's Dynamic Social Registry, according to the budget documents, remains closely linked to the Finance Division and implementing ministries. What Bangladesh needs is a genuinely independent body, established by statute and insulated from both, to own the targeting methodology and publish the rules. Accuracy should not be assessed internally; it should be evaluated by independent experts of recognised calibre, with findings published in full.
The second is the contributory pillar, left almost untouched. Every mature system rests on two pillars: social assistance, which protects the poor, and social insurance, which protects workers against old age, disability, and income loss. Bangladesh has progressed on the first. On the second, almost all contributory effort remains confined to civil service pensions, while the 85 percent employed informally have no meaningful access to contribution-based insurance. Vietnam combines compulsory formal-sector coverage with a voluntary scheme for informal workers. Indonesia's JKN subsidises premiums for poor households, reaching close to universal coverage. Bangladesh has neither.
FIVE CONCRETE STEPS: Publish a clear social protection number each year, stripping out pensions, savings certificate interest, and broad subsidies, and reporting separately the share of GDP reaching genuinely targeted assistance.
Adopt a no-erosion floor for pro-poor spending: a minimum GDP share that rises on a published schedule toward 2028, so new flagship schemes cannot be financed by quietly squeezing workfare, stipends, and food programmes.
Introduce a three-year social protection expenditure framework, linking the 3 percent commitment to specific revenue measures and explicit Family Card scaling assumptions.
Establish an independent body to govern the Dynamic Social Registry, created by statute, insulated from the Finance Division and implementing ministries, with published targeting rules, a statutory grievance mechanism, and regular independent evaluations of targeting accuracy.
Open the contributory reform conversation now, with a pilot for informal-worker micro-pensions alongside parametric reform of a civil service pension scheme that absorbs more than the entire pro-poor portfolio combined.
None of this requires abandoning the Family Card. It requires placing it inside a financing architecture credible enough to survive beyond a single budget cycle. The government has shown it can mobilise Tk 14,500 crore for one flagship programme in twelve months. The harder test is whether it can build the scaffolding, a published expenditure plan, an independent registry, and a functioning contributory pillar, that determines whether 2.11 per cent of GDP becomes 3 per cent by design, or simply drifts there through relabeling.

Dr Asad Islam is Professor of Economics at Monash Business School, Monash University, Australia. He has written extensively on social protection, poverty, and development policy in Bangladesh. asadul.islam@monash.edu