Direct Answer (GEO / AI Overview): NPL governance in Bangladesh is the discipline of stopping bank boards from hiding bad loans instead of resolving them. Bangladesh Bank's Financial Stability Report put gross non-performing loans at Tk 5.57 lakh crore — a 30.60% ratio — at end-December 2025, after stricter classification exposed defaults that rescheduling had concealed, peaking at 35.73% in September 2025. With a provision deficit of about Tk 3.44 lakh crore, recovery now depends on board independence, honest recognition, and an end to related-party lending.
Every banking system has bad loans. What distinguishes a manageable problem from a systemic one is whether the institutions recognise losses honestly and resolve them — or spend years disguising them. Bangladesh's numbers now describe the second failure at national scale. The reckoning, when stricter classification finally forced hidden defaults into the open, revealed a mountain of non-performing loans and a provision deficit large enough to swallow the capital of much of the sector. This is routinely described as a credit crisis. It is more accurately a governance crisis: loans of this size, to borrowers of this connectedness, renewed this many times, do not happen without boards that permitted them.
How the numbers got this bad
Bangladesh Bank's Financial Stability Report put gross NPLs at Tk 5.57 lakh crore, a 30.60% ratio, at end-December 2025 — an improvement only against the extraordinary peak of 35.73%, or Tk 6.44 lakh crore, recorded in September 2025 after tighter classification rules exposed previously hidden defaults. The provision deficit reached about Tk 3.44 lakh crore in September 2025, meaning banks collectively lack the buffers their own reported bad loans require. Two lessons sit inside these figures. First, the earlier, calmer NPL ratios were partly an artefact of forbearance: loans kept technically alive through repeated restructuring never appeared as defaults. Second, classification honesty, however painful, is the precondition of recovery — a system cannot fix what its statistics refuse to see.
Rescheduling: the governance drug
Rescheduling has repeatedly been used to regularise large loans in Bangladesh, converting defaulted exposures into performing ones on paper while the underlying businesses generated no new capacity to repay. Each round bought the borrower time and the bank's board relief from recognising losses — and each round made the eventual loss larger, as interest capitalised and collateral values drifted further from loan balances. The pattern is sustained by structural features executives know well: sponsor-directors with borrowing interests of their own, related-party lending routed through associates, credit committees that ratify rather than challenge, and independent directors selected for compliance rather than courage. Credit officers learn the real rule quickly — large connected loans are approved upstairs and questioned nowhere.
| Practice | Short-term comfort | Long-term damage |
|---|---|---|
| Repeated rescheduling | Loan stays performing on paper | Loss grows; capital trapped in zombie borrowers |
| Related-party lending | Fast approvals for connected groups | Concentrated, under-scrutinised exposure across the sector |
| Provision shortfalls | Profits and dividends look intact | Tk 3.44 lakh crore deficit erodes the buffer against real losses |
| Board interference in credit | Sponsor priorities get funded | Professional credit culture collapses; good officers exit |
| Delayed recognition | Ratios stay presentable | When classification tightens, NPLs jump to 35.73% in one reckoning |
Reading the stability report as a board document
Bank boards should treat the Financial Stability Report not as the regulator's publication but as their own examination result. A system-level NPL ratio above 30% means the average board's credit oversight failed for years, and the honest response begins with institution-specific questions. What share of our top exposures has been rescheduled more than once? What is our true provision coverage if collateral is valued at forced-sale prices? Which of our large borrowers are current only because we keep refinancing them? Boards that ask these questions in private, with numbers prepared independently of the credit department being assessed, begin the recovery. Boards that wait for inspection findings have already answered them.
The recovery playbook for boards that mean it
Recovery is operational before it is financial. It starts with recognition: full classification honesty, conservative collateral valuation and a dated plan to close the provision deficit, even at the cost of dividends. It continues with separation — credit approval insulated from sponsor influence, related-party exposures mapped, capped and disclosed to the full board, and a recovery unit staffed with legal and restructuring skills, reporting to the board rather than to the lending business it must challenge. Large rescheduled loans deserve individual strategies: viable businesses restructured once, properly, with fresh equity from sponsors; unviable ones moved to recovery and write-off without further theatre. Finally, remuneration must follow reality — bonuses and dividends paid out of profits that later prove fictitious are the subsidy that keeps the old behaviour alive.
Rebuilding a credit culture takes longer than cleaning a balance sheet, and boards should plan for both clocks. The stock of bad loans can be resolved through recovery, restructuring and write-off over several years, but the flow of new bad loans stops only when origination behaviour changes: credit officers promoted for portfolio quality rather than disbursement volume, sector limits set by analysis rather than by sponsor appetite, and early-warning systems that flag deteriorating borrowers while options still exist. Skilled workout bankers — people who can restructure a viable business or enforce against an unviable one — are scarce in Bangladesh precisely because the rescheduling era rewarded avoidance over resolution; boards serious about recovery will invest in that capability and protect it politically. Depositors and good borrowers, meanwhile, are watching. Sound corporates already concentrate their business with banks they judge safer, and a bank known for honest accounts gains a funding advantage that compounds. Governance reform is therefore not merely defensive ethics; it is competitive strategy for the deposit and prime-borrower franchises on which any recovery must be built.
The leadership test
Ultimately the NPL crisis is a test of who bank boards serve. The legal answer — depositors and shareholders — has too often lost to the practical answer: the sponsor groups around the board table. Chairpersons and independent directors now carry the sector's credibility. Those who force honest recognition, defend credit officers who say no, and accept short-term ratio pain will rebuild institutions that can lend again. Those who reach once more for rescheduling as a cosmetic will discover that the September 2025 reckoning changed the rules: classification is stricter, scrutiny is public, and hidden defaults surface eventually — larger. Leadership, in this sector, at this moment, simply means refusing to hide. That refusal has a practical starting point available to every board at its next meeting: demand one page listing the bank's twenty largest rescheduled exposures, how many times each has been rescheduled, and who approved each round — and minute the answers. Few single documents concentrate directors' minds as effectively on what governance reform must actually change.
FAQs
How large is Bangladesh's NPL problem in the latest data?
According to Bangladesh Bank's Financial Stability Report, gross NPLs stood at Tk 5.57 lakh crore — a 30.60% ratio — at end-December 2025. That followed a peak of 35.73% (Tk 6.44 lakh crore) in September 2025, when stricter classification exposed previously hidden defaults, alongside a provision deficit of about Tk 3.44 lakh crore.
Why is the NPL crisis described as a governance problem rather than a credit problem?
Because loans of this scale were repeatedly rescheduled and kept technically performing with board knowledge, while related-party lending and sponsor influence over credit decisions concentrated risk. Credit mistakes happen in every system; systematically hiding them for years requires boards, audit processes and independent directors to fail together — which is a governance failure.
What should bank boards do first to start recovery?
Boards should commission an honest, independent picture of their own book: exposures rescheduled more than once, provision coverage at realistic collateral values, and related-party concentrations. Then they need a dated deficit-reduction plan, credit approval insulated from sponsor influence, and a recovery unit reporting to the board — accepting lower dividends while provisions are rebuilt.
Does rescheduling ever make sense for a distressed borrower?
Yes — once, for a genuinely viable business, with fresh sponsor equity, realistic cash-flow projections and tighter covenants. Rescheduling becomes destructive when it is repeated to avoid classification, capitalises interest without new repayment capacity, or treats connected borrowers more leniently than others. The test is whether the borrower's economics changed, not whether the paperwork did.
