Why digital nano-loans are reshaping Bangladesh fintech

The City Bank–bKash digital nano-loan platform crossed Tk 10,000 crore in cumulative disbursements in July 2026, covering more than 3.19 crore individual loans taken by approximately 3.5 million bKash customers.

The scale matters less as a headline than as evidence of a structural shift. Credit that previously required a branch visit, formal documentation, collateral or an established banking relationship can now be assessed and disbursed through a mobile financial services account. The model does not eliminate credit risk; it changes how that risk is measured, priced and distributed across Bangladesh's financial system.

The central development is therefore not simply that bKash users can borrow small sums. It is that transaction data, regulated banking capital and mobile distribution have been combined into a lending channel capable of operating at a scale that conventional microcredit and retail banking could not easily match.

The evolution of collateral-free credit in Bangladesh

The first fully digital, collateral-free nano-loan platform in Bangladesh was launched by City Bank and bKash on December 15, 2021. The commercial launch followed a year-long pilot supervised by Bangladesh Bank, which began in July 2020.

That sequence established the institutional architecture of the product. bKash provides the digital interface and customer distribution network, while City Bank provides the lending facility. bKash is not the licensed bank issuing loans from its own balance sheet. The credit is disbursed by the partner financial institution through a system integrated into the bKash application.

This distinction is material. Bangladesh's mobile financial services sector has become a major channel for payments, remittances and account access, but the statutory framework for lending remains connected to licensed banks and financial institutions. The digital platform simplifies the customer journey; it does not remove the underlying regulatory division between a mobile money operator and a bank.

The product was built around a narrow but commercially important credit gap. Many consumers and micro-entrepreneurs require amounts too small, too urgent or too irregular for traditional retail lending processes. A shopkeeper may need short-term liquidity to restock inventory. A salaried worker may face a timing mismatch between household expenses and income. A mobile user may need a limited amount to settle a payment obligation rather than finance a long-term asset.

For these borrowers, the transaction cost of formal credit can be disproportionate to the loan amount. A branch-based process designed for larger loans is not economically efficient when the requested sum is BDT 500 or BDT 3,000. Digital nano-lending addresses that imbalance by lowering origination costs and making repayment part of the same mobile ecosystem through which the loan was requested.

The available loan range on the City Bank–bKash platform is between BDT 500 and BDT 50,000. The average disbursement is around BDT 3,500, indicating that the platform is primarily serving short-duration liquidity needs rather than conventional investment or asset-financing requirements.

Digital nano-loans are not a smaller version of conventional banking; they are a different operating model built around high-frequency data and low transaction costs.

The platform's reported scale is significant in relation to the size of individual loans. It has disbursed approximately 100,000 loans per day and around Tk 900 crore per month. A system processing this volume cannot rely on manual underwriting as its primary control mechanism. Its viability depends on automated eligibility decisions, standardized pricing and rapid repayment data.

How the bKash–City Bank lending model works

The lending model uses artificial intelligence-based credit scoring and the customer's bKash transaction history to determine eligibility and the available credit limit. The amount is not uniform across the user base, and access is not automatic for every bKash customer.

The stated eligible customer pool exceeds 1.2 crore bKash users, but eligibility within that pool depends on individual transaction patterns and the scoring model. In practical terms, the platform is assessing behavioural evidence rather than relying exclusively on conventional collateral or extensive documentary income verification.

The relevant signals may include the consistency of account activity, the frequency and type of transactions, repayment behaviour and the stability of the customer's digital financial usage. The precise composition of the scoring model is not public, but its underlying logic is clear: a customer's transaction history becomes a proxy for financial regularity and repayment capacity.

This is a major departure from conventional credit assessment in markets where many potential borrowers have limited formal credit histories. The digital account is not merely a payment instrument. It becomes a source of underwriting data.

The model also changes the economics of loan servicing. When the average loan is approximately BDT 3,500, the cost of manually reviewing applications, collecting documents and maintaining physical records would consume a material share of potential revenue. Automated processing allows the bank to serve low-value loans while preserving the possibility of portfolio-level profitability.

The pricing structure reflects a regulated lending product rather than an informal mobile application loan. The interest rate is 9% per annum, accrued on a daily basis, while the processing fee is 0.575% of the total loan amount, comprising a 0.5% base fee plus VAT.

For a borrower, the nominal annual interest rate is only one part of the cost. The effective cost depends on the amount borrowed, the period for which it remains outstanding, the processing fee and the repayment schedule. The 0.575% processing fee is proportional to the loan amount and therefore scales with the principal rather than acting as a flat charge. The total cost to the borrower is determined by the interaction of tenure and the daily interest that accrues over the life of the loan. A short-term facility may accumulate limited interest in absolute terms, but the longer a balance remains outstanding, the more the daily rate compounds into the total repayment figure.

That pricing structure is one reason digital credit should be analysed through cash-flow mechanics rather than headline interest rates alone. A BDT 500 facility and a BDT 50,000 facility may carry the same stated annual rate, but the relative impact of fees, timing and repayment discipline will differ substantially.

The platform's operational design can be summarised through its core variables:

• Loan size: BDT 500 to BDT 50,000, with an average disbursement of approximately BDT 3,500.
• Credit decision: Determined through AI-based scoring and the customer's transaction history.
• Distribution channel: The bKash mobile application, without a branch visit or physical paperwork.
• Lending institution: City Bank, operating as the regulated financial provider.
• Interest rate: 9% per annum, calculated on a daily basis.
• Processing fee: 0.575% of the loan amount, including VAT.
• Risk outcome: Reported default levels below 1% on the platform.

The technical layer is important, but the institutional arrangement is more important. Digital interfaces can accelerate loan applications; only disciplined underwriting, enforceable repayment processes and adequate bank-level controls can sustain a lending portfolio.

From nano-loans to Pay-Later finance

The next stage of the market has been the extension of digital credit from cash disbursement to point-of-sale finance. In April 2024, bKash and City Bank introduced Pay-Later, allowing customers to complete purchases through bKash merchant QR codes and repay either within seven days without interest or through six-month instalment plans.

This is a shift in the use case. A conventional nano-loan places liquidity in the customer's account, leaving the borrower to decide how to spend it. Pay-Later connects credit directly to a merchant transaction. That gives the lender more information about the purpose and timing of the borrowing while embedding repayment into the digital payment relationship.

The product also supports the expansion of merchant acceptance. If consumers can use a mobile wallet to access short-term purchasing power, merchants gain an additional payment mechanism that does not depend entirely on cash availability at the point of sale. The resulting effect is not necessarily a broad increase in consumption; it is a change in the settlement infrastructure through which everyday purchases are financed.

The distinction between productive and consumptive borrowing remains relevant. Small business users may employ short-term credit to purchase stock or smooth working capital. Households may use it to manage temporary income volatility. Other users may use Pay-Later facilities to bring forward purchases that would otherwise have been delayed.

The same platform can support all three behaviours. That makes portfolio monitoring more complex. Aggregate repayment performance may remain strong even while the underlying borrower purposes differ significantly.

The rise of digital credit also needs to be viewed alongside wider changes in household finance. Higher interest rates and tighter financial conditions can alter the cost of borrowing across formal and informal channels, as outlined in this analysis of how elevated interest rates reshape personal finance. For Bangladesh's digital lenders, the issue is not simply whether customers want instant credit, but whether their income and cash-flow conditions can absorb repeated short-term obligations.

A digital lending product can reduce the friction of borrowing. It cannot eliminate the borrower's repayment constraint.

Bangladesh Bank and the formalisation of e-loans

The regulatory direction has moved toward formal recognition of fully digital lending. In May 2026, Bangladesh Bank issued circular guidelines allowing scheduled banks to disburse fully digital e-loans of up to BDT 50,000, with tenures of up to 12 months, without branch visits or physical paperwork.

This policy expands the significance of the bKash–City Bank model beyond one partnership. It creates a broader regulatory pathway through which commercial banks can develop digital credit products within defined limits.

The BDT 50,000 ceiling is particularly relevant because it corresponds to the upper limit of the existing nano-loan platform. The framework therefore provides a regulatory basis for products that are small enough to be automated but large enough to support household liquidity, micro-enterprise working capital and short-term commercial payments.

The maximum tenure of 12 months also indicates that Bangladesh Bank is not treating every digital loan as an ultra-short emergency facility. Banks may use the framework for a range of repayment structures, provided that underwriting, documentation and reporting are handled digitally within the approved statutory framework.

The regulatory approach has several implications.

First, it should reduce the incentive for borrowers to rely on unauthorised mobile applications that operate outside the formal banking system. The Bangladesh Financial Intelligence Unit has warned against unauthorised mobile app-based lending, a risk that becomes more serious when digital access is mistaken for regulatory legitimacy.

Second, it allows Bangladesh Bank to bring digital lending into a more consistent supervisory perimeter. Products that are originated, serviced and repaid through mobile applications still generate credit exposure. Their digital format does not make them exempt from prudential concerns involving consumer protection, data use, classification of loans and operational resilience.

Third, it may increase competition among scheduled banks. Banks that previously lacked an efficient way to reach low-value borrowers can now use digital channels and transaction-based underwriting. The competitive advantage will not rest only on pricing. It will depend on access to distribution, quality of data, model governance and the ability to control fraud.

The formalisation of e-loans is therefore a market-structure development. It may bring more lenders into the segment, but it will also place greater pressure on banks to demonstrate that automated decisions are explainable, monitored and consistent with responsible lending requirements.

Why the default rate remains below 1%

The reported default level below 1% on the City Bank–bKash platform is one of the strongest indicators of the model's current performance. It is also the figure most likely to be misunderstood.

A low default rate does not mean that digital nano-lending is inherently low risk. It means that this particular combination of customer selection, loan limits, repayment design, transaction data and regulated oversight has produced a controlled portfolio outcome.

The comparison with unregulated digital lending markets, including default rates of up to 40% reported in Kenya, illustrates the importance of institutional design. A mobile application alone does not create a viable lending system. Where lenders depend on aggressive acquisition, weak identity controls or opaque collection practices, the portfolio can deteriorate rapidly.

The Bangladesh model benefits from several forms of risk containment:

1. Small initial exposure limits losses. A borrower may receive as little as BDT 500, while the maximum facility is capped at BDT 50,000. The lender can therefore increase exposure incrementally rather than beginning with a large unsecured balance.

2. Transaction history improves screening. The scoring system uses observed activity within the bKash ecosystem, creating a behavioural basis for eligibility decisions where traditional credit records may be incomplete.

3. The distribution channel is familiar to customers. Borrowing, receiving funds and making repayments occur through an established mobile financial services interface rather than an unfamiliar standalone app.

4. The banking partner carries regulated responsibility. City Bank's role places the lending product within a formal financial institution, with corresponding requirements for controls, oversight and portfolio management.

5. Digital repayment reduces collection friction. The lender can communicate with customers and process repayments through the same channel used for disbursement, reducing the operational burden associated with cash collection.

These factors are mutually reinforcing, but none is permanent. A scoring model can weaken if customer behaviour changes, if the eligible pool expands too quickly or if borrowers take multiple facilities across different platforms. A default rate below 1% at an established scale is an important result; it is not a guarantee that expansion will preserve the same performance.

The central risk is portfolio migration. Once a customer becomes accustomed to instant credit, repeated borrowing may replace one-off liquidity management. If wage growth, business turnover or household cash flows weaken, repayment performance can deteriorate even when the original credit decision was sound.

There is also a data governance issue. Transaction-based lending depends on the collection and interpretation of behavioural information. The more financial decisions are automated, the greater the importance of model governance, access controls and transparent treatment of customers whose data does not fit the system's preferred patterns.

Digital inclusion is not identical to universal approval. A platform that rejects some users because their transaction history is insufficient may be operating prudently, even if that exclusion is commercially inconvenient. The policy challenge is to expand access with safeguards, not to loosen them. The same technology that allows a BDT 500 loan to reach a tea-stall owner within minutes also allows that loan to be the first of many. The institutions that built this market — bKash, City Bank and the regulators now formalising the rules — will be judged less on the size of cumulative disbursements than on whether the next 3.19 crore loans perform as well as the first.