The confluence of an IMF programme, a newly elected government with a parliamentary supermajority, record remittance inflows of $30.33 billion in FY2024-25, and the most significant regulatory overhaul of the financial sector in the country’s history has created a window for structural reform that may not reopen for another generation. The question is no longer whether reform is necessary; that debate was settled when Bangladesh Bank’s Asset Quality Reviews revealed NPLs of 35.73 percent of total disbursed loans by late 2025, equivalent to over Tk 6.44 trillion in bad assets. The question now is sequencing, political will, and technical execution.
The Bangladesh Bank’s adoption of Basel III loan classification standards, under which a loan becomes sub-standard after just three months of overdue installments, caused the official NPL ratio to jump from 10.11 percent in June 2023 to 20.20 percent by December 2024, and to an unprecedented 35.73 percent by late 2025. This is not a deterioration. This is a revelation. The deterioration happened over the preceding decade, hidden beneath a veneer of politically driven loan rescheduling, cosmetic provisioning, and what Finance Adviser Dr Salehuddin Ahmed correctly described as “rampant embezzlement.” Distressed assets are likely two and a half to three times the reported NPL figure once off-balance-sheet exposures, restructured loans extended to connected parties, and under-provisioned sovereign-adjacent exposures are properly accounted for.
The governance failure was structural and bipartisan. State-owned banks, holding less than 30 percent of banking assets, account for more than 45 percent of problem loans. Several private banks, particularly Islamic banks with concentrated exposure to a few designated groups, were essentially hollowed out. Bangladesh Bank’s response from August 2024 onwards: board reconstitution, forced mergers, liquidity injections, and the Islamic bank consolidation were without precedent. But the asset recovery task force has produced limited results. The stolen money is largely abroad. And the new BB exposure rules, whereby raising single-borrower limits to 25 percent of capital, risk reinforcing exactly the concentration dynamics that produced this crisis.
On the fiscal side, Bangladesh’s tax-to-GDP ratio has hovered at 7-8 percent for years, one of the lowest in the developing world. The country is graduating from LDC status in 2026, which means the trade preferences that subsidised the apparel sector’s export competitiveness will begin to erode. Corporate tax for unlisted firms stands at 27.5 percent, above the regional average of approximately 21 percent. Tax compliance takes approximately 435 hours annually. Revenue forecasting credibility is so low that NBR missed its July-November FY2026 target by Tk 24,000 crore, even while posting 15 percent year-on-year growth.
Meanwhile, FDI stood at $1.77 billion in 2025, still below the 2019 peak of $1.8 billion; concentrated in textiles, finance, and power, with limited diversification into the higher-value sectors Bangladesh needs for upper-middle-income status by 2031.
This is the platform on which we must build.
Phase 1: Complete the diagnostic and draw the line (FY2026 – immediate)
The single most important thing Bangladesh Bank can do right now is complete the Asset Quality Reviews for all state-owned banks before the June 2026 deadline it has set for itself. Independent AQRs by firms of the calibre of KPMG and EY, already conducted for private banks, must be applied to Sonali, Janata, Agrani, and Rupali with equal rigour and equivalent public disclosure. The numbers will be ugly. That is precisely the point. You cannot build a recovery plan on numbers that nobody believes.
The NPL reduction targets that BB has imposed: state-owned banks to 10 percent, private banks below 5 percent by June 2026, are the right signal, but are almost certainly unachievable in the given timeframe for the state banks. The response to missing these targets should not be to adjust the targets. It should be to accelerate the resolution tools and be honest with the public about the timeline. Bangladesh Bank’s credibility, which was badly damaged under the previous dispensation, is rebuilt only through candour, not through adjusted benchmarks.
BB’s new Prompt Corrective Action (PCA) framework, triggered by capital shortfall, NPL thresholds, and dividend policy violations, must be applied without exception. The new dividend restrictions, under which no cash dividends can be paid if the capital adequacy ratio falls between 10 and 12.5 percent, are exactly right. The temptation to grant discretionary waivers to politically connected bank owners under the new government must be resisted as firmly as it was not resisted under the last.
Phase 2: Build the resolution architecture (FY2026–2027)
The Bank Resolution Ordinance 2025 is the most consequential piece of financial legislation Bangladesh has enacted since the Bangladesh Bank Order 1972. It gives BB the legal authority to intervene in failing banks without waiting for court proceedings, create bridge banks, protect depositors, and limit judicial challenges to compensation claims rather than the resolution action itself. This is the framework. The substance must now follow.
Distressed Asset Management Act: This legislation, when enacted, must create a genuine secondary market for NPL trading, not a mechanism for connected parties to purchase distressed assets at discounts and return them to affiliated borrowers at par. The asset management companies (AMCs) licensed under this framework must have foreign participation requirements to prevent the regulatory capture that doomed Bangladesh’s previous attempts at distressed asset resolution. The decision not to create a state-funded “bad bank” is correct; public money should not absorb private sector fraud.
Bankruptcy Act: Bangladesh desperately needs a corporate insolvency framework that allows genuinely distressed but viable companies to restructure under creditor supervision, rather than defaulting entirely and dragging bank balance sheets down with them. India’s Insolvency and Bankruptcy Code 2016, for all its implementation imperfections, has demonstrated that a credible insolvency regime changes creditor-debtor behaviour ex ante, not just ex post. Bangladesh’s version must provide a genuine time-bound resolution process, 180 days with extensions, rather than the de facto open-ended litigation that the Money Loan Courts currently produce.
Money Loan Court Reform: The pledge to amend the Money Loan Court Act by the first quarter of FY2026 is welcome. But amendments alone are insufficient. Bangladesh Bank must fund a dedicated case management unit within the court system, staffed with transaction advisors, asset valuers, and enforcement specialists, as an independent agency, not a committee of generalist civil servants. Recovery performance is currently measured by the number of cases filed, not the money recovered. This perverse incentive must be reversed.
The “Golden Exit” and “Willful Defaulter” Distinction: BB’s current framework, allowing loan rescheduling with a two percent down payment for genuine businesses, is pragmatic. But the critical reform is the legal definition and enforcement of “willful default”, where borrowers have the capacity to repay but choose not to. The asset recovery task force, working with BFIU, ACC, and international partners, must pursue these cases publicly, internationally, and to conviction. One high-profile recovery from an international jurisdiction does more for credit culture than a hundred circulars.
Phase 3: Structural reform of the state bank model (FY2027–2029)
The Policy Research Institute has proposed the most analytically coherent framework: privatise all state banks except Sonali, which should be restructured as a pure treasury operations bank. This is directionally correct but politically difficult. A more sequenced approach is achievable:
Janata, Agrani, and Rupali banks should undergo full corporatisation with stock market divestment, management control transferred to professional boards without ministry interference, and clear performance contracts signed with the Bangladesh Bank rather than the Finance Division. The dual oversight arrangement, whereby state banks report to the Finance Ministry rather than to the central bank, is one of the most damaging structural features of Bangladesh’s banking system. Basel III compliance cannot coexist with ministry-directed lending.
Sonali Bank retains strategic importance for government payments infrastructure and rural access. It should be restructured as a narrow bank authorised for deposits and government treasury functions, with a ring-fence around its NPL portfolio under a dedicated resolution vehicle.
Basic Bank and Bangladesh Development Bank have no viable commercial case. Supervised wind-down under the Resolution Ordinance is the appropriate end-state, with depositor protection through an enhanced Deposit Insurance Scheme.
Phase 4: Building the preventive architecture (FY2026–2030 ongoing)
The proposed amendment to the Bangladesh Bank Order 1972 to grant full operational and regulatory independence is foundational to every other reform. All reforms above are reversible if BB can be instructed by the Finance Ministry or PMO. The amendment must include: fixed-term appointments for the Governor and Deputy Governors removable only for cause; a statutory prohibition on ministerial direction on supervisory decisions; and independent funding through a levy on supervised institutions rather than the Treasury.
IFRS 9 (ECL) implementation by December 2027 is the single most important accounting reform, with risk-based supervision. Under the current incurred-loss provisioning model, banks provision only when a loan has already gone bad. Under IFRS 9’s Expected Credit Loss model, banks must provision based on the forward-looking probability of default. This will force additional capital raising, cause further short-term NPL spikes, and place enormous pressure on weaker banks. The transition must be accompanied by a well-designed capital market access plan so that solvent, but under-capitalised banks can raise tier-1 capital through public issuance.
The Monetary Policy Statement for H1 FY2026 correctly identifies over-reliance on bank financing as one of Bangladesh’s critical structural vulnerabilities. The bond market reform agenda: Green Sukuk, listing of large domestic companies, commodity exchanges, and blockchain-based settlement must be accelerated. Banks must not continue to be asked to finance the entire private sector capital structure. A functioning bond market absorbs long-tenor infrastructure risk from bank balance sheets, extends corporate maturities, and provides an alternative savings instrument to time deposits.
The Revenue Policy and Revenue Management Ordinance of May 2025, separating the former NBR into a Revenue Policy Division and a Revenue Management Division, was structurally correct and long overdue. The separation of policy from administration is a global best practice, implemented successfully across OECD countries precisely because it eliminates the conflict of interest where the body writing tax law is also the body collecting tax and can informally negotiate its own compliance targets.
The six-week officer strike and the subsequent partial retreat by the interim government do not invalidate the reform architecture: it reveals the institutional resistance that must be managed rather than appeased.
The medium-to-long-term revenue strategy (MLTRS FY2026-FY2035) correctly identifies that Bangladesh could reach a 15 percent tax-to-GDP ratio without raising rates if compliance improves and leakages are eliminated. Modelling by PRI suggests annual revenue shortfalls of approximately Tk589 billion from the current system alone. The roadmap to 12 percent by 2030 requires five simultaneous interventions:
The integration of the Customs Modernisation Strategic Action Plan, ASYCUDA World, online income tax filing, VAT e-invoicing, and the proposed national taxpayer identity system must be treated as a single program under a central programme management office, not as a collection of individual departmental IT projects. Bangladesh Bank’s real-time payment infrastructure (RTGS, BEFTN) already exists; connecting it to the Revenue Management Division’s collection systems would close a major leakage point instantly.
Bangladesh’s VAT system has multiple rate slabs, sector-specific exemptions, and supplementary duty structures with approximately 113 variations in tax incidence, a figure that contradicts WTO principles and creates enormous compliance arbitrage. The reform must reduce the number of effective VAT rates to no more than three (zero, a reduced rate for essential goods, and the standard rate), extend the VAT base to the currently-exempt service sector, and replace most supplementary duties with a simple, transparent excise system. This does not require raising rates; it requires closing the gap between the statutory base and the effective base. VAT performed relatively better than income tax in recent collection data precisely because it is harder to evade at the point of sale. Extending this logic to services, where most of Bangladesh’s formal sector value-add now resides, is the single largest available tax base expansion.
Bangladesh has approximately 3.5 million registered taxpayers against a workforce of 70 million and an economy of $460 billion. The gap is structural, not incidental. Three targeted interventions close most of it: mandatory TIN registration for all formal sector employees and directors (already underway but unenforced), cross-matching of bank transaction data with declared income (Bangladesh Bank and the revenue authority must share data under a legal framework that prevents misuse), and a graduated presumptive tax for the informal sector that is low enough to encourage voluntary compliance rather than avoidance. The fear of “tax terrorism” is real; the revenue authority’s historical use of discretionary assessment powers as an extortion mechanism is well-documented. This is precisely why separating policy from enforcement and making the assessment process fully digital and therefore auditable is a precondition rather than a nice-to-have.
Several current Bangladesh Bank prudential guidelines, while individually defensible, collectively create an investment climate that is harder to navigate than it needs to be. Three specific changes would make an immediate difference:
The current restrictions on retained earnings repatriation, the complexity of the back-to-back LC framework for non-garment industries, and the USD/BDT swap market’s shallow liquidity are among the most frequently cited barriers by foreign investors. BB should establish a Foreign Investor Facilitation Desk with delegated authority to resolve individual investor FX issues within 10 working days, similar to Singapore’s MAS One-Stop framework. The broader reform is a managed shift toward current account convertibility under a rules-based managed float, with a transparent intervention framework. The era of informal BB intervention in the interbank FX market, which created the parallel rate differential that contributed to remittance diversion, must end with a clear policy statement.
The policy rate is currently held at 10 percent, with the Standing Lending Facility at 11.5 percent. This is appropriate given that inflation, while declining, remains above the 6.5 percent June 2026 target. However, the transmission mechanism from the policy rate to lending rates remains weak, particularly for SME credit. The Quarterly Financial Stability Report should include a mandatory section on monetary policy transmission metrics by bank category, forcing banks to explain divergences between lending rates and the policy corridor.
Bangladesh’s commercial banks lend predominantly to large corporates, as the BB lending limit discussion above confirms. SMEs, which generate most of Bangladesh’s employment, access credit at punitive rates or not at all. BB should establish a Credit Guarantee Fund, capitalised through a small levy on large corporate lending, that provides first-loss coverage for SME loans below Tk5 crore meeting defined performance criteria. Bangladesh Bank’s existing green finance taxonomy should be expanded and linked to the National Determined Contributions under the Paris Agreement, with risk-weighted capital relief for qualifying green exposures as an incentive for bank lending into the sector.
A corporate tax rate of 27.5 percent for unlisted firms, against a regional average of approximately 21 percent, is a structural competitive disadvantage at a moment when Bangladesh needs to attract manufacturing investment from China-plus-one supply chain diversification and compete with Vietnam, Cambodia, and Indonesia for the next tier of export industries. The rate reduction should be sequenced against broadening: lower the rate to 22.5 percent in FY2027, to 20 percent in FY2029, conditional on demonstrable improvements in compliance enforcement and the phase-out of sector-specific exemptions. A lower headline rate on a broader base raises more revenue than a high headline rate on a narrow, riddled base.
The gap between listed (22.5 percent) and unlisted (27.5 percent) corporate tax rates exists to incentivise stock exchange listing. It should be maintained and even widened slightly to 8 percentage points as part of a deliberate capital market deepening strategy. Forcing large domestic firms to list, which Bangladesh Bank’s Monetary Policy Statement for H1 FY2026 endorses, will not succeed through exhortation. The tax incentive, combined with a functional regulatory framework for IPOs, is the mechanism.
Bangladesh Investment Development Authority (BIDA) has been institutionally strengthened but remains operationally fragmented. A foreign investor seeking to establish a manufacturing facility must navigate BIDA, BEZA (for economic zone allocation), NBR (for tax incentive registration), BB (for foreign exchange approvals), and the relevant line ministry for sector-specific licenses, each with independent documentation requirements, processing times, and informal expectations.
The solution is not another “one-stop shop” circular; Bangladesh has had several of those. The solution is a genuine integrated digital processing system, with BIDA as the lead agency, legally binding processing timelines (not aspirational ones), and, critically, a senior-level escalation mechanism that involves the PM’s office directly when processing stalls. Investors in Vietnam and Indonesia know that if the investment bureaucracy fails to respond in 30 days, they can call a government hotline and expect a response from a Deputy Minister. That accountability infrastructure does not yet exist in Bangladesh, and it must be built.
The IMF programme, backed by the World Bank and ADB, has provided the external scaffolding for these reforms. The frustration with delays in meeting conditions, which reportedly led to the withholding of $800 million in disbursements pending the new government, is a warning, not a detail. Bangladesh’s access to concessional financing, critical for bridging the LDC graduation fiscal gap, depends directly on reform performance.
The sequencing logic should be as follows. In FY2026, the priorities are completing the banking sector AQRs, submitting the NPL resolution legislation to parliament, legislating the NBR restructuring through parliament rather than an ordinance, and publishing the first full Tax Expenditure Report. These are all largely institutional and legislative actions that do not require fiscal space; they require political will and administrative capacity.
In FY2027, the execution phase begins: IFRS 9 transition underway, first AMC licensed, Bankruptcy Act operational, first corporate tax rate reduction implemented, and the digital tax-banking data bridge operational.
By FY2029, the sustainable architecture should be in place: BB fully autonomous, state bank privatisation substantially complete, tax-to-GDP at 10 percent, and a functional corporate bond market absorbing long-tenor investment financing that currently sits on bank balance sheets.
None of this is technically difficult. Every instrument described above has been implemented in comparable middle-income economies. What has been missing in Bangladesh is the combination of political insulation from vested interests, institutional memory inside the civil service, and sequenced execution discipline.
Bangladesh’s financial sector failed its people not through market forces but through deliberate regulatory capture. The roadmap is visible. The current reform moment, with IMF accountability, donor support, new political leadership, and a central bank that has demonstrated it can act independently, is the best opportunity in a generation to build institutions that outlast any single political dispensation.
History will not remember who managed the first ECNEC meeting, but it will most certainly remember who fixed the banks at a time of complete distress.
