Bangladesh's banking sector is confronting an uncomfortable reality.
Overdue agricultural credit across all scheduled banks reached Tk 22,915 crore in February 2026 -- a staggering 124% increase from Tk 10,234 crore in the same month of 2025.
The headlines are alarming, and the instinct in many financial boardrooms will be to pull back from rural lending entirely.
As agricultural overdues rise, conventional wisdom suggests that lending to smallholder farmers is a high-risk proposition plagued by low recovery rates.
However, a deeper analysis reveals a different story: The perceived "risk" was never about the farmers’ intent; it was a failure of traditional lending architecture to provide visibility and structured cash-flow linkages.
Traditional agricultural lending has long operated with limited visibility into how funds are used, how crops perform, and how cash flows materialize.
Without this visibility, even well-intentioned lending becomes difficult to manage, leading to rising delinquencies.
The other side of the coin is that every year, Bangladesh Bank allocates a reasonable amount as a target for agriculture lending.
During the 2024-25 financial year, banks disbursed Tk 32,211 crore in agricultural loans, achieving 84.76% of the total disbursement target for FY25, according to Bangladesh Bank data.
For the 2025-26 fiscal year, Bangladesh Bank has set a record agricultural credit disbursement target of Tk39,000 crore.
The invisible farmer problem
For decades, Bangladeshi smallholders have remained financially invisible. Without documented transaction histories, formal land records, or credit trails, they are simply unreadable to conventional banking systems.
But there is a counter-narrative unfolding quietly in the rural farms of Bangladesh.
There has been a rise of agritech companies since 2019, who are essentially trying to solve the challenges within the agriculture value chain, using technology.
But only a few of them have been showing early traction and trust amongst the farmers. Certain companies have been fundamentally rewriting the agriculture lending playbook in Bangladesh -- and their business model treats agricultural lending not as charity, but as a systemic problem.
The results speak for themselves: a 97% loan recovery rate, maintained consistently even as the broader sector deteriorates. More than 14,000 farmers are currently integrated into the formal financial system through this platform, backed by banking partners.
There is scope for others to follow suit by adopting some of the measures such as using machine learning to analyze farmers' behavioural patterns, farming practices, and historical performance to build a credible creditworthiness profile.
Field agents use GPS mapping to verify farm boundaries, while remote sensing technology monitors crop health in real time using NDVI indices, allowing banks to confirm that what is on the loan application actually exists in the ground.
Agents physically accompany farmers to digitize their financial histories and validate records against government databases. The farmer becomes legible. The risk becomes quantifiable.
A proof of concept the sector needs
The macro picture remains sobering. Bangladesh Bank has said the surge in overdues signals growing concerns over credit recovery and the need for stronger oversight and risk management in the sector.
The regulator is right. But stronger oversight does not mean less lending, it means smarter lending infrastructure.
To avoid worsening the overdue crisis, the banking sector will need to transition from collateral-based lending to data-driven, ecosystem-backed models.
Farmers are not the risk. The real challenge is the infrastructure gap, and it has a solution.
Fahad Ifaz is co-founder and CEO of iFarmer.