Back to Research
Narrative 2026-05-16

Stuck at Eight Percent

Bangladesh has not raised its tax-to-GDP ratio in three decades. Six months from LDC graduation, that floor has become the ceiling.

Stuck at Eight Percent

Executive Summary

Bangladesh's tax-to-GDP ratio stood at 9.0 percent in 1990. In 2024, it stood at 8.3 percent. Across thirty-four years of growth, governments, and constitutional change, the fiscal floor barely moved. As LDC graduation in November 2026 eliminates the trade preferences and concessional borrowing terms that quietly subsidised this under-collection, the country faces a direct revenue reckoning. This brief identifies three levers that can close the gap: property tax modernisation, VAT base recovery, and a five-year sunset program on tax expenditure, together worth up to 3.1 percentage points of GDP by 2030.

On a Tuesday in May 2026, a sub-registrar's office in Gulshan completes paperwork for a 2,400-square-foot flat that has just changed hands. The seller and the buyer agree, off the record, that the actual price was Tk 8 crore. The deed value entered on the registration paper is Tk 1 crore 40 lakh, drawn from the mouza rate that the government has not refreshed for the building's plot in seven years. Stamp duty, registration fee, gain tax, source tax, local government tax: all of them are calculated on Tk 1.4 crore. The state collects roughly Tk 14 lakh on a transaction whose real magnitude was almost six times larger. The remaining Tk 6 crore 60 lakh of capital gain leaves the system as cleanly registered, freshly white wealth.

Two hundred and fifty kilometres north, the head teacher of a single-shift government primary school in Sherpur sets out three pieces of chalk for the day's classes. The annual non-salary line in her school's contingency budget runs to a few thousand taka, which has to cover chalk, blackboard markers, sweeping supplies, exam paper, and minor repairs. She buys the rest out of her own pocket. The reason the budget is not ten times larger is the same reason there is no working ventilator at the upazila health complex, no functioning agricultural extension office in three of her district's eight unions, and no asphalt left on the road her students walk in the rains.

The country is not poor. The country does not collect.

A flat line for thirty-four years

Bangladesh's tax revenue, measured as a share of GDP, was 9.0 percent in 1990. According to IMF Article IV data, the same ratio was 8.3 percent in 2024. The IMF's 2025 Article IV consultation, concluded in January 2026, observed that tax-to-GDP fell sharply in FY25 alongside a decelerating economy: real GDP growth slowed to 3.7 percent in FY25 from 4.2 percent in FY24 and 5.8 percent in FY23, reflecting production disruptions during the 2024 popular uprising, a tighter policy mix, and sluggish investment. NBR's FY25 collection totalled Tk 3.69 trillion. In between the 1990 starting point and today, the ratio dipped under 7 percent in 2001 and 2017, recovered to 8.5 percent in 2022, and is forecast at 8.9 percent for 2026. Across thirty-four years, four governments, three constitutional amendments, two military caretakers, one major sovereign rating cycle, and a roughly twelve-fold increase in real GDP, the tax floor moved by less than a percentage point.

This is not a story about Bangladesh being poor. Vietnam crossed our 1990 level of tax effort within a decade of starting its own reform period and is now collecting somewhere between sixteen and nineteen percent of GDP depending on the year. Thailand collected 15.3 percent in 2024. Indonesia, a country whose tax-to-GDP is itself considered low for its income bracket, sits around eleven. Even India, whose central-government tax revenue at 6.7 percent of GDP in 2022 is below ours, brings in another ten or so percentage points through state-level taxes, putting consolidated Indian tax effort near seventeen. The Bangladesh number is not the South Asian normal. It is the South Asian floor.

The honest framing is this: every reform that costs money is bottlenecked by this ratio. Bangladesh spends about two percent of GDP on education, against a global lower-middle-income average above four. It spends below one percent of GDP on public health, against a WHO floor recommendation of around five percent of GDP combined public and private. Interest payments on existing debt now absorb above twenty percent of total government expenditure, while health absorbs about nine. The constraint is not what we want to do. The constraint is what we can pay for.

What is collected, and what is not

When you look inside the NBR's roughly Tk 2,76,000 crore collected in FY2024, the composition tells the story. VAT is 39 percent. Income tax is 38 percent. Customs duty is 11.8 percent. Supplementary duty is 10.6 percent. Excise is half a percent.

Of that income tax block, the share that comes from individuals filing personal returns is small. The country has fewer than four million active TIN-holders who file returns in a given year, in a population of about 175 million and a labour force above 70 million. Corporate income tax dominates. Personal income tax acts more like a deduction at source on salaries in the formal sector than a system of voluntary annual filing by individuals with wealth.

VAT is the harder failure. Bangladesh's value-added tax system was modernised on paper through the VAT and Supplementary Duty Act of 2012, fully implemented in stages from 2019. The promise was a broad-based, electronically administered consumption tax that would raise the VAT share of GDP toward four percent over a decade. The opposite happened. VAT collection as a share of GDP fell from 2.83 percent in FY2013 to 1.92 percent in FY2022 and has only partially recovered to 2.17 percent in FY2024. C-efficiency, which is the standard diagnostic ratio of actual VAT collected to what a perfectly enforced VAT on observed consumption would yield, fell from 18.9 percent in FY2013 to a low of 12.8 percent in FY2022 before edging back to 14.5 percent in FY2024. A VAT system that becomes less efficient every year for a decade is not a tax system; it is a leak.

The third component is customs and supplementary duty, the trade tax block, which still funds 22 percent of total NBR collection. That share has been falling for fifteen years, from 32.8 percent in FY2010, and that decline is good news as a sign of a more domestically grounded tax base. It is also a slow-burn fiscal vulnerability, because the post-LDC tariff schedule that Bangladesh will face after November 2026 will only reinforce that erosion. The country's own published roadmap implicitly accepts that trade taxation will keep shrinking.

What is not collected is more telling than what is. Property is not taxed in any meaningful way. The municipal holding tax in Dhaka, Chittagong, and the other corporations is tied to legacy assessed rental values that have nothing to do with current market values, and even those rolls are decades out of date. The government's own tax expenditure statement, published by the NBR for FY2024, estimates revenue forgone through exemptions and incentives at around three percent of GDP, which is more than the entirety of VAT collection in the same year. Capital gains on land and apartment transfers are routinely declared at mouza rate, the government's notified per-decimal value, which in central Dhaka is commonly under one-fifth of the transacted market price. Inheritance tax does not exist. Wealth taxation, beyond the symbolic surcharge on net wealth above a high threshold, does not exist.

A state that does not tax wealth, does not tax land, does not tax intra-family transfers, and cannot keep its VAT efficient is not a state that has reached the limits of what its economy can bear. It is a state that has chosen, year after year, who to ask for money and who to leave alone.

The November cliff

On November 24, 2026, six months from this Tuesday morning in Gulshan, Bangladesh will graduate from the United Nations' Least Developed Country category. The graduation itself is a credit. The economic side effects are a bill.

The European Union's Everything But Arms preference, which gives our garments duty-free access to the world's second-largest single market, expires three years after graduation. After that, default tariffs of nine to twelve percent on RMG products kick in unless we have negotiated a successor arrangement, most likely under the EU's GSP+ scheme. GSP+ requires ratification and credible implementation of twenty-seven international conventions covering labour rights, environmental standards, human rights, and good governance. Several of these conventions, in particular those affecting freedom of association in export processing zones, have been blocked by industry interests for the better part of two decades.

The TRIPS pharmaceutical waiver, which allows our generic drug industry to manufacture without holding the relevant patents, lapses on a slightly different timetable. The DCTS framework with the United Kingdom is more accommodating but covers a smaller export share. Concessional sovereign borrowing terms tighten: IDA classification erodes, IBRD pricing applies, and the cost of new external borrowing rises by something on the order of two hundred basis points on the marginal dollar.

Pre-computed estimates in the bdpolicylab fiscal results put the customs and supplementary-duty revenue at risk from LDC graduation at between 10 and 25 percent of the current trade-tax base over the medium term. In taka terms, that is a loss of somewhere between Tk 6,200 crore and Tk 15,400 crore, equivalent to between 2.2 and 5.6 percent of total NBR collection at the FY2024 level. The mid-case figure of about Tk 11,000 crore is roughly equal to the entire annual development budget for the Department of Agricultural Extension or the operational budget of the Ministry of Health and Family Welfare's non-salary line.

The point of citing these numbers is not to argue against graduation. It is to argue that the country has to replace this revenue from somewhere, and the answer cannot be more deficit financing through Bangladesh Bank or more 9-percent T-bill issuance to the same banking system that is already absorbing too much of the deficit and crowding out private credit. The answer has to be domestic revenue mobilisation. There is no third door.

The three levers that actually move the needle

The bdpolicylab reform roadmap models three sequential phases that, if executed, lift tax-to-GDP from 8.3 percent in 2024 to 9.4 percent by 2026, 10.4 percent by 2028, and 11.4 percent by 2030, the full 3.1-point gain from the three levers. The 2030 endpoint would still leave Bangladesh below Thailand's 2024 level. Even modest convergence requires three specific moves that no government for thirty-four years has been willing to make.

First, property tax modernisation at the local-government level. The mechanical fix is administrative, not political in the deep sense: mouza rates have to be revised to reflect current transacted prices, with a transparent index and a five-year refresh cycle written into law rather than left to discretionary updating. City corporation holding tax rolls have to be rebuilt with current market valuations and made publicly searchable, the way they are in most Indian metros. Capital gains on real estate have to be calculated on the higher of declared and indexed market value, with the index published. The full revenue potential is large: a serious property tax regime in the four metropolitan corporations alone could plausibly add 0.5 to 0.8 percent of GDP within five years. The political cost is concentrated on the urban real estate developer class and on second-generation political families whose primary wealth sits in land.

Second, VAT base broadening and efficiency recovery. The C-efficiency target is to climb from 14.5 percent back to the 19 percent range Bangladesh briefly achieved in 2013, which on current consumption would add about 0.8 percent of GDP. The administrative fix is mandatory point-of-sale issuance and online integration for retailers above a low turnover threshold, replacing the current system where the registration threshold is high and policed loosely. The policy fix is to end the proliferation of reduced rates and product-specific exemptions added through repeated Finance Acts, which have made the supposed standard 15 percent rate apply to less than half of true consumption. The political cost is concentrated on mid-sized retail, the wholesale market sector, and the industrial lobby that has bargained for product-specific exemptions over the past decade.

Third, sunsetting the tax expenditure sprawl. The NBR's own FY2024 tax expenditure statement quantifies foregone revenue at roughly three percent of GDP. A serious five-year sunset program, with each exemption requiring a public-interest justification to be renewed and a default expiry if no justification is filed, would over time recover one to one and a half percentage points of GDP. The exemptions cover textiles, special economic zones, IT and IT-enabled services, agriculture inputs, capital machinery imports, listed corporate dividends, and dozens of smaller categories that were originally introduced as time-bound incentives and were never allowed to expire. The political cost is concentrated, directly, on the industries that have built their margins around the exemption.

Together these three levers, if executed in parallel rather than sequentially, would close roughly half the gap to the LDC peer median by 2030. They do not require any new policy idea that has not been studied. They require a government willing to take the political hit of taxing wealth, taxing consumption properly, and ending exemption capture.

The reform that opens every other reform

Everything that follows in this series, on banking sector reform, on energy transition, on education quality, on climate adaptation, on the long arc of demographic-dividend exhaustion, runs into the same wall: there is no money. The state does not collect enough to recapitalise the banks honestly. The state does not collect enough to subsidise the solar tender pipeline at the scale industrial users need. The state does not collect enough to pay teachers what would clear the labour market for competent recruits. The state does not collect enough to fund coastal managed retreat. Every one of these problems is solvable. None of them is solvable at 8 percent of GDP.

In November 2026, the country will lose access to the trade preferences and concessional borrowing terms that have quietly subsidised the under-collection of taxes for the better part of two decades. After that, the deficit becomes the constraint, the deficit becomes the lender of last resort, and the lender of last resort is Bangladesh Bank, and Bangladesh Bank in 2010 to 2024 already showed what happens when a central bank becomes the fiscal authority's printer.

The Tarique Rahman BNP government, sworn in on February 17, has roughly twelve months from its inauguration to push a credible revenue package through parliament. The shape of the package is already known: property tax modernisation, VAT base recovery, exemption sunset. The data is already collected. The numbers are already published. The only missing input is the willingness to ask the people who can pay to pay.

Sources

  • IMF Article IV 2025 (concluded January 2026): https://www.imf.org/en/news/articles/2026/01/30/pr-26029
  • NBR FY25 tax-to-GDP decline: https://thefinancialexpress.com.bd/economy/bangladesh/nbrs-fy25-tax-to-gdp-ratio-falls
  • NBR Medium- and Long-Term Revenue Strategy FY2025-FY2035: https://nbr.gov.bd/uploads/publications/MLTRS_E_Book_Version.pdf
  • World Bank tax revenue indicator (Bangladesh): https://data.worldbank.org/indicator/GC.TAX.TOTL.GD.ZS?locations=BD
  • BDPolicyLab fiscal model: precomputed LDC-graduation and reform-lever scenario outputs

Three pieces of chalk for an entire morning is not what this country looks like. It is what this country has chosen to look like, by leaving Tk 6 crore 60 lakh in the back pocket of a Gulshan flat sale every Tuesday for the last fifteen years.

Created: 2026-05-16 14:39:21 Updated: 2026-05-29 19:43:11