The State of Bangladesh Banking: Crisis, Reform, and Resilience

Chapter 1

The Banking Landscape

Bangladesh's banking sector comprises 61 scheduled banks across four ownership categories: 6 state-owned commercial banks (SOBs), 3 specialized development banks, 43 private commercial banks, and 9 foreign banks. This structure, a legacy of post-independence nationalization and successive private licensing waves in the 1980s and 1990s, creates an oversaturated market where too many banks compete for a limited pool of bankable assets. The resulting fragmentation depresses profitability, encourages connected lending, and complicates regulatory oversight. The Sammilito Islami Bank merger in late 2025, which absorbed five failed banks, reduced the effective count by four and set a precedent for state-led resolution.

The officially reported non-performing loan ratio rose from 1.9% in 2011 to 9.6% in 2023 under the pre-review reporting regime. That trajectory already reflected deep governance failures: directed lending by SOBs to politically connected borrowers, weak credit appraisal at private banks, and Bangladesh Bank's historically constrained enforcement capacity. Asset-quality reviews and Basel III classification then corrected years of forbearance-masked reporting: classified loans rose to 20.20% of total loans by December 2024, 24.1% by March 2025, and a record 35.73% by late 2025 (Bangladesh Bank, 2026, reported by CPD). Little in the underlying portfolio changed across these revisions; the classification finally caught up with it.

Lending rates moderated from over 12% in the early 2000s to around 9.9% by 2024, partly shaped by Bangladesh Bank's 9% lending rate cap (imposed 2020, replaced by SMART rate 2023). Compressed margins undermine banks' ability to absorb credit losses and build capital buffers, a cost that compounds as NPLs accumulate. The trade-off between cheap credit access and sector solvency is now the central pricing question for the new central bank leadership.

The Sammilito merger changes the structure in one direction only: it removes failed private banks from the count and places their liabilities on the state balance sheet. Whether it improves system solvency depends on the quality of the asset-management and recovery process that follows. The architecture of resolution now exists; the enforcement discretion has changed hands.

Chapter 2

Credit, Money, and Interest Rates

Private sector credit as a share of GDP expanded from about 22% in 2000 to 35.8% by 2024, reflecting the economy's deepening financial intermediation. Yet this ratio sits well below peer economies such as Vietnam (147%) or Thailand (148%), indicating that Bangladesh's formal banking sector still plays a limited role in channeling savings to productive investment. A large share of economic activity in agriculture, micro-enterprises, and the informal sector remains outside the formal credit system, and the NPL-driven credit contraction of 2024-2025 is widening that gap further.

Broad money (M2) as a share of GDP has grown from under 14% in 1974 to 48.8% in 2024. M2 growth has moderated sharply in recent years (6.1% in 2024) as Bangladesh Bank tightened monetary policy to combat double-digit inflation and as credit demand contracted under the weight of NPL uncertainty. The central bank's shift to interest-rate-based monetary policy in mid-2023 represents a structural change in framework, though transmission remains constrained by administered rates and weak monetary pass-through in a sector dominated by connected lending.

Interest Rate Dynamics

The lending-deposit spread has narrowed from over 4 percentage points in 2000 to 1.3 pp in 2024. A narrower spread benefits borrowers on paper but compresses bank margins at a time when provisioning demands for rising NPLs require greater earnings capacity. The new BB Governor has signaled a preference for lower policy rates to revive private credit growth; that choice interacts directly with the sector's ability to self-provision against the 35.73% classified loan stock.

The real lending rate has turned negative during high-inflation episodes, meaning depositors effectively subsidize borrowers. This distortion discourages financial savings, accelerates capital flight, and diverts household resources toward real assets, all outcomes that weaken the deposit base the banking cleanup depends on.

Chapter 3

Regional Benchmarking

Comparing Bangladesh against regional peers reveals a sector that lags on nearly every dimension of financial health. Even on the pre-review reported figure, Bangladesh's NPL ratio of 9.6% already led the comparison group: Pakistan 6.6%, Thailand 2.8%, Vietnam 2.3%, Indonesia 2.0%, India 1.7%. A like-for-like comparison against the late-2025 Bangladesh figure is not exact, because peers report standard NPL ratios while the 35.73% Bangladesh number is a post-review classified-loans ratio; the gap is wide on any reasonable adjustment. India improved dramatically through the Insolvency and Bankruptcy Code; Bangladesh has the legal precedent but not yet the institutional follow-through.

Financial depth tells the same story. Bangladesh's credit-to-GDP ratio of 35.8% is one of the lowest in the comparison set. Thailand leads at 148%, underscoring the gap between Bangladesh's banking sector and what is achievable at similar or higher income levels. The M2-to-GDP ratio of 48.8% is respectable in isolation but masks the fact that much of the monetary expansion has been absorbed by government borrowing rather than channeled to private productive investment.

The five-dimension radar synthesizes NPL quality, credit depth, lending rate, M2/GDP, and branch density. Bangladesh trails the peer average on all but M2/GDP, and the gap on asset quality widened materially once the asset-quality review replaced the reported series with the honest one.

The peer comparison does one thing the domestic data cannot: it provides a credible floor for what reform could deliver. India cut its NPL ratio from over 11% in 2018 to 1.7% in 2023 through statutory resolution. Thailand and Vietnam maintain ratios below 3% while supporting faster credit growth than Bangladesh. The gap is not structural; it is institutional.

Chapter 4

Stress and Vulnerability

Foreign exchange reserves bottomed at end-2024: gross reserves of $26.21 billion, or $21.39 billion on the IMF BPM6 standard (Bangladesh Bank Financial Stability Report 2024). Against a monthly import bill of roughly $5.5 billion (Bangladesh Bank, Dec 2025), that BPM6 level covered about 3.9 months of imports. The position has since recovered substantially: gross reserves stood at $34.32 billion on May 19, 2026, and $29.67 billion on the BPM6 measure (BSS), held up by sustained remittances and import compression. The BDT/USD rate settled around 122 after the managed depreciation of 2022-2023, a correction that was overdue but imposed significant costs on importers and dollar-denominated borrowers.

The most acute fiscal risk lies in the state-owned banks. Conservative estimates place the total recapitalization need for SOBs at approximately 555 BDT billion (roughly $4.5 billion at current exchange rates), close to 3% of GDP as a one-time stock. This is a BDPolicyLab estimate built from SOB capital shortfalls under Basel III minimum capital adequacy; it is not a sum of published per-bank recapitalization plans, which the government has not released bank by bank. Without recapitalization, SOBs cannot meet Basel III minimum capital ratios, constraining their lending capacity and eroding depositor confidence. The government faces a choice between absorbing the fiscal cost and allowing zombie banks to continue operating with negative net worth.

The banking-sector capital hole is connected to the energy-sector capital hole: independent power producer capacity charges flow through state-owned bank loan portfolios, and unpaid power-sector liabilities sit on BPDB's balance sheet, which the state stands behind. None of these can be resolved in isolation. The fiscal envelope set by the Tarique Rahman government budget constrains all of them simultaneously.

Chapter 5

Reform Scenarios and Outlook

Three scenarios frame the divergent paths Bangladesh's banking sector could take through 2030. The status quo baseline starts not from the pre-review 9.6% but from the honest 35.73% classified-loans level. Under no meaningful change to regulatory enforcement or SOB governance, that ratio stagnates: new NPLs are generated at roughly the same rate as old ones are written off or rescheduled, producing no net improvement in asset quality. The Sammilito merger is already complete; whether it becomes a one-off or a template depends on political will under the new BB leadership.

A moderate reform path, one that includes strengthened loan classification standards, independent SOB boards, and a functional asset management company, could bring the NPL ratio toward 6.0% by 2030. The aggressive reform scenario, modeled on India's IBC experience, envisions full Basel III compliance, time-bound NPL resolution, privatization of at least two SOBs, and credible license revocation for persistently undercapitalized banks. This path could reduce NPLs to 3.0% by 2030.

The reform sequencing is critical. BB governance reform, board restructuring, and operational autonomy codified in statute (not personal undertakings) must come first. An asset management company, strengthened insolvency legislation, and forensic audit publication for every bank above 20% NPL form the core 2025-2026 reform package. SOB restructuring and mobile financial services regulation extend the horizon into 2027 and beyond. Governor Mostaqur Rahman's public commitment not to bow to political pressure is the entry condition; the institutional lock-in that survives his successor is what matters.

The Tarique Rahman government has spoken about asset recovery in general terms. It has not yet named the first beneficiary against whom a recovery case will be filed. The window in which prosecution is politically possible narrows as new business relationships form. Without asset recovery, the precedent is set that capture has no cost beyond the loss of operational control of the franchise.

Policy Implications

Toward a Resilient Banking Sector

The analysis across five chapters points to a sector that is functional but fragile, capable of supporting Bangladesh's current growth trajectory but ill-equipped for the demands of an upper-middle-income economy. The recommendations below name an owner, an expected effect, and a measurable signal that the action is working.

1. Convert the Sammilito liability into a working balance sheet. Owner: Bangladesh Bank and Ministry of Finance. The merger of five failed Shariah banks into Sammilito Islami Bank PLC is complete (Tk 35,000 crore paid-up capital, of which Tk 20,000 crore is a government injection from the FY2025-26 budget, about 0.7% of GDP). The outstanding task is converting the depositor liability into a sustainable balance sheet through an asset-management vehicle and a published recovery timeline. Success signal: net recoveries against transferred bad assets exceed 10% of book value within 18 months, reported quarterly.
2. Phase the SOB capital shortfall over five years. Owner: Ministry of Finance, with Bangladesh Bank supervision. Recapitalize or restructure state-owned banks against time-bound performance contracts. The estimated 555 BDT billion shortfall (about 3% of GDP as a stock) is large but manageable phased at roughly 0.3% of GDP annually. Success signal: every SOB meets the Basel III minimum capital adequacy ratio by 2030, with no further forbearance extensions granted in the interim.
3. Legislate Bangladesh Bank independence into statute. Owner: Cabinet and Parliament. The new governor's verbal commitment not to bow to political pressure does not survive a change of governor. Amend the Bangladesh Bank Order so the operational independence of the rule-enforcement function is irreversible, not contingent on the governor of the day. Success signal: a statutory amendment passed that sets fixed-term protection and a for-cause-only removal standard for the governor.
4. Establish a national asset management company. Owner: Ministry of Finance. A centralized AMC, modeled on Korea's KAMCO or Malaysia's Danaharta, can purchase and resolve distressed assets at scale, freeing bank balance sheets for new lending. The Sammilito resolution provides the institutional learning for design. Success signal: enabling legislation enacted and the AMC capitalized and operational within 24 months.
5. Publish bank-level forensic audit findings. Owner: Bangladesh Bank. Extend the forensic audits that preceded the Sammilito merger to every commercial bank with a classified-loan ratio above 20%, and publish the findings. Without the underlying numbers in the public domain, the recapitalization debate at the next budget is uninformed and political-economy pressure to soften terms is unconstrained. Success signal: audit summaries for all above-threshold banks published before the FY2026-27 budget.
6. File the first named asset-recovery case. Owner: Bangladesh Bank, Anti-Corruption Commission, and the Attorney General. The precedent that capture costs nothing beyond loss of operational control holds until a recovery case is filed against a named beneficiary. Success signal: at least one high-value recovery case filed and publicly docketed within the current government's first year, with frozen assets reported.

Methodology and Sources

Data and Attribution

Quantitative results in this report draw on pre-computed data in data/banking_flagship_results.json (BDPolicyLab internal), which aggregates Bangladesh Bank annual reports and Financial Stability Reports, World Bank World Development Indicators, IMF Financial Soundness Indicators, and Bangladesh Bureau of Statistics. Exchange rate data are from the Bangladesh Bank interbank market. The post-review classified-loans series (Dec 2024, Mar 2025, late 2025) is sourced to Bangladesh Bank (2026) as reported by the Centre for Policy Dialogue and is not in the pre-computed results JSON, which ends at Dec 2023. Import-cover months are computed as BPM6 reserves divided by the Bangladesh Bank monthly import bill of about USD 5.5 billion (Dec 2025). The SOB recapitalization figure is a BDPolicyLab estimate of the aggregate Basel III capital shortfall, not a sum of published per-bank plans.

• Classified loans 35.73% of total loans, late 2025 (Bangladesh Bank, 2026) (CPD, State of the Bangladesh Economy in FY2025-26 (First Reading)): cpd.org.bd/state-of-the-bangladesh-economy-in-fy2025-26-first-reading
• Sammilito Islami Bank: Tk 35,000 cr paid-up capital, Tk 20,000 cr govt injection (The Daily Star): thedailystar.net/business/economy/banks/news/sammilito-islami-bank-gets-bb-nod-4031156
• Govt removes BB governor Ahsan H Mansur, Mostaqur Rahman appointed (Feb 25, 2026) (New Age): newagebd.net/post/Banking/292237/govt-removes-bb-governor-ahsan-h-mansur-mostaqur-rahman-appointed
• Who is Mostaqur Rahman, the new cenbank governor (Hera Sweaters, BNP committee) (The Business Standard): tbsnews.net/bangladesh/who-mostaqur-rahman-new-cenbank-governor-1371051
• Shock ouster of BB governor (The Daily Star): thedailystar.net/news/bangladesh/news/shock-ouster-bb-governor-4115256
• End-2024 reserves: gross USD 26.21 bn, BPM6 USD 21.39 bn (Bangladesh Bank, Financial Stability Report 2024): bb.org.bd/en/index.php/publication/publictn/3/10
• Forex reserves gross USD 34.32 bn, BPM6 USD 29.67 bn (May 19, 2026) (BSS): bssnews.net/business/388633
• Internal data (BDPolicyLab): data/banking_flagship_results.json