ILLUSTRATION: DANIEL JURMAN
IN THE spring of 2022, land in Thamel, Kathmandu’s commercial district, was selling for around Rs30m per anna, about 3.5 square metres. Ten years earlier, the same land would have cost roughly one-fifth of that. Nepal’s banks treated this sharp rise in prices as security rather than a risk. They lent against the land, refinanced it and built much of the financial system on its rising value. The bill is now arriving.
According to figures published by the Nepal Rastra Bank in January, 88.3% of banks’ total loans—Rs4.55trn out of Rs5.15trn outstanding—are backed by land or buildings. That single number changes how the rest of the system should be viewed. Banks have not been engaged in broad-based lending. Instead they have tied most of their balance-sheets to property, building a highly leveraged exposure to Nepal’s real-estate market while presenting it as a varied lending business.
Also read: Kathmandu’s property prices have decoupled from reality
The consequences are now bleeding through the accounts. Non-performing loans were 1.41% of total lending in mid-2021, a period still influenced by the aftermath of the covid-19 pandemic. By January this year they had leapt to 5.26%, a fourfold increase. There has been no single banking scandal or sudden external shock in that time—only the gradual unwinding of an overextended credit cycle as pandemic-era support faded and property prices cooled.
Among Class C finance companies, which sit at the most vulnerable end of Nepal’s three-tier banking structure, the non-performing loan ratio has hit 11.86%. Roughly one loan in eight at these institutions will not be repaid in full.
The origins of the crisis trace back to a confluence that, with hindsight, looks almost engineered for disaster. Remittances have rocketed from about 15% of GDP in 2010 to more than 25% today, as large numbers of Nepalis working in the Gulf, Malaysia and Japan send money home each month. Much of it ends up as bank deposits in Kathmandu and in towns across the country, creating a steady and inexpensive source of funds for lenders.
In an economy with a small stockmarket dominated by financial firms, no retail bond market and savings that have struggled to outpace inflation, there were few productive outlets for this money. Land and property took the bulk of it.
A single transaction shows how the system works. A family in Sindhuli receives Rs100,000 from a son in Qatar and deposits it at a local commercial bank. With few creditworthy borrowers in agriculture or manufacturing nearby, the bank channels part of that deposit into a Rs50m construction loan for a developer in Kathmandu. The developer builds apartments, which are bought on mortgage by upper-middle-class families in the capital, many of whom also receive remittances. Rural savings have thus been converted into urban property speculation. The process was gradual, widespread and largely invisible.
The regulatory framework failed to halt it. Nepal Rastra Bank’s capital adequacy rules give relatively low risk weightings to property-backed loans, on the assumption that collateral provides security. The reasoning is circular. Collateral only protects the bank if the property can be sold near its appraised value when a borrower defaults. Defaults tend to cluster in downturns, the very periods when property cannot fetch boom-era prices. The safeguards were designed for the scenario in which they were least necessary and absent when they mattered most.
The credit-to-deposit ratio illustrates the cycle with brutal clarity. In January 2022, at the height of the boom, banks were lending 95p for every rupee deposited, close to the regulatory limit while non-performing loans sat at a tranquil 1.18%. By January 2026 the ratio had fallen to 74% as banks pull back from new lending, while NPLs quadrupled. This is the textbook pattern of a credit cycle: the boom creates the bust, and the bust exposes the risks the boom had hidden.
What went unnoticed was a basic confusion between collateral and creditworthiness. During the boom, Nepal’s banks judged borrowers mainly by the value of the assets they pledged, not by their actual ability to generate income. A Kathmandu developer with fully mortgaged land could borrow to build apartments, assuming the sales would cover the loan—so long as prices rose as expected, in a market the banks themselves were inflating. A household receiving irregular remittances from the Gulf could take on a mortgage consuming most of its income, on the assumption that the remittances would continue indefinitely at the same level. Banks were not checking whether borrowers could repay. They were checking whether the property could be sold to cover the debt if repayment failed. They were wrong about both.
A secondary structural flaw made the valuation problem worse. Nepal lacks a centralised property transaction database. Land prices are privately negotiated but officially recorded at values that consistently understate actual sale prices, largely to reduce transfer taxes. Approved valuers, reliant on banks for work, had strong incentives to produce appraisals that would allow loans to proceed. Properties were repeatedly revalued upward at each refinancing, letting borrowers draw fresh credit against assets that had created no new economic value. When the cycle turned and banks sought to sell collateral from defaulted loans, they found that forced-sale prices in Kathmandu bore little relation to the book values on which their lending had been based.
The provisioning data offers the clearest sign that bank management realises the problem is worse than the published NPL ratios suggest. Loan-loss provisions for the six months to January reached Rs57.0bn, up 48.6% from the same period the year before. Banks do not boost provisions lightly: doing so reduces reported profits, depresses dividends and invites awkward questions from shareholders. The surge is, in effect, the sector’s internal credit committees acknowledging that loans classified as performing today will not be performing tomorrow. The gap between published NPLs and the true state of the loan book is being closed silently, one provision at a time.
The gap between banks is striking. Nabil Bank, regarded as one of Nepal’s more disciplined lenders, set aside provisions equal to just 4.6% of its interest income and posted a net profit of Rs4.76bn for the half-year. Laxmi Sunrise Bank, which completed a 2022 merger combining balance sheets that had not been fully stress-tested, booked provisions at 47.2% of interest income and recorded a net loss of Rs273.6m—the only commercial bank in the sector to do so. NIC Asia Bank, with one of the highest immediate exposures to the property sector, earned a net profit of Rs131.2m on a Rs415bn balance sheet. That margin will vanish if provisioning requirements rise further, as its NPL trajectory suggests they will.
Among the Class C finance companies, the situation is even more advanced. These 17 institutions occupy a difficult position: too small to compete with commercial banks for prime borrowers, yet too regulated to lend informally, and heavily concentrated in consumer credit and property—the segments driving most of the NPL rise. They expanded during the boom by offering slightly higher deposit rates to attract funds, then lent more aggressively than commercial banks would. Their reported capital adequacy ratio of 13.0% may overstate their strength if loan classifications lag the true level of impairment. Without either a property market recovery or recapitalisation, further deterioration appears inevitable.
he deposit side of the crisis shows a paradox in the wider dynamics. Total deposits reached Rs6.95trn in January, up 28% over two years, as remittance flows continued despite the banking sector’s difficulties. Lending by contrast grew only 16% over the same period. The gap between deposit growth and lending reflects banks’ reluctance to take on new credit risk amid high NPLs and uncertain collateral values. The surplus is parked in government securities and central bank instruments. Nepal’s banks are thus simultaneously overflowing with deposits and too cautious to lend them—maintaining stable net interest margins while contributing little to the economy’s productive capacity.
The families sending remittances into this system are not, in any meaningful sense, investors in Kathmandu property. They are workers on Gulf construction sites and in Malaysian factories who placed their savings where they thought them safest. Their deposits—earning 2.93% a year—helped fund the property cycle. Nepal has no deposit insurance scheme capable of handling systemic stress; the Deposit and Credit Guarantee Corporation’s coverage is limited, and its fund is not large enough to cover failures at the institution level. If resolution proceedings become necessary for Class C finance companies or a mid-tier commercial bank, retail depositors will face uncertainty rather than assured recovery. The risk was created in Kathmandu’s property market and will fall, at least in part, on the remittance accounts of Nepalis working abroad.
The cost of this misallocation is evident not only in rising NPLs but in the lending that never happened. Agricultural credit, supporting a sector that employs most of Nepal’s population, stood at Rs314.2bn in January, down from Rs330.1bn in mid-2023. In real terms bank lending to farming is shrinking. Nepal’s diverse climatic zones—capable of producing everything from large cardamom and orthodox tea to high-altitude herbs and speciality coffee—have received only a fraction of the credit that Kathmandu apartment developments attracted. A coffee farm in Gulmi provides no land title that can be readily appraised; a plot in Thamel does. The choice was made accordingly, year after year, branch by branch, loan committee by loan committee.
Hydropower likewise paints a nasty picture. Nepal sits atop an estimated 80,000 megawatts of technically exploitable capacity, of which less than 3,500 MW has been developed. Lending to the electricity, gas and water sectors has grown at a healthy 18.3% year on year and is a real bright spot in the portfolio. Yet at Rs461bn, it represents just 9.0% of total lending. Nepal’s infrastructure-focused NIFRA, the dedicated project finance institution, has a balance sheet of around Rs42bn—smaller than many commercial banks’ individual loan portfolios. The gap between Nepal’s development needs and the banking sector’s allocation choices could hardly be clearer.
Three scenarios now emerge. In the most favourable, property prices stabilise; rescheduled borrowers gradually reduce their obligations; and NPLs plateau below 7% before falling from 2027 onwards. This would require no major remittance shock, continued government infrastructure spending, and no further interest rate pressure—a combination of positive developments that has rarely occurred simultaneously.
More likely, according to analysts closely monitoring the sector, is a prolonged workout: NPLs could rise to 7–9% over the next two years, several Class C institutions may need resolution through merger or orderly wind-down, and commercial bank profitability could be constrained for three to five years as provisioning consumes most interest income. The sector would emerge smaller and less profitable but solvent, provided the Nepal Rastra Bank manages the process actively rather than relying on the cycle to correct itself.
The tail risk is a shock to remittances. Gulf labour markets are not fixed. A significant slowdown in hiring in Qatar, the UAE or Saudi Arabia—whether from falling oil prices, geopolitical disruption or automation in construction and logistics—could cut deposit inflows by 20–30% within a year or two. A banking sector already burdened with high NPLs and falling credit-to-deposit ratios would face simultaneous pressure on assets and funding. Nepal’s bank resolution framework has not been tested on this scale, and depositor confidence, currently stable, would become uncertain.
None of this was inevitable. A banking sector genuinely focused on Nepal’s development might have lent based on verified cash flows rather than collateral values, created agricultural finance instruments that accounted for seasonal and climatic risks instead of avoiding them, and developed the long-term project finance expertise hydropower requires. It would have needed regulators prepared to limit property concentration through counter-cyclical capital requirements, and credit cultures willing to undertake the harder task of assessing borrower capacity rather than the easier task of valuing land titles. Those choices were not made. The sector that exists today is the result of the choices that were.
The bill is now being shared. Shareholders who took dividends during the boom are seeing returns shrink. Borrowers who stretched to buy at peak prices are defaulting. Depositors, who sought only the safety of their funds, are discovering that the assets behind those deposits are worth less than the accounts suggested. Meanwhile, the wider Nepali economy continues to lack the productive credit that could have supported farms, factories and power plants under a development-focused banking system. The land titles were valued. The opportunity cost was not. ■
