The country's financial system is enormous on paper. But numbers don't repay loans - people do, and right now, people aren't sure tomorrow is safe.

Nepal's banks are sitting on Rs 7.7 trillion in deposits. Total credit outstanding is pushing Rs 5.8 trillion. As a share of GDP, those numbers would make some middle-income countries jealous, deposit penetration at 126%, credit at nearly 95%. By those metrics, Nepal has a financially "deep" economy.

And yet something is clearly wrong.

The non-performing loan ratio across commercial banks, development banks, and finance companies is 5.42%, already past the threshold that European banking supervisors use as a formal trigger for tighter scrutiny. Finance companies are at 11.86%. That's not a yellow flag. That's a red one. The loan-loss provisions being held against potential defaults now equal 5.83% of total lending, which sounds reassuring until you realize what it means: banks are quietly setting aside money they expect not to get back, and that provisioning is eating into the earnings they'd otherwise use to build capital.

A banking system can look big and broken at the same time. Nepal's currently manages both.

What actually went wrong is less about the size of the lending book and more about what the lending book is made of.

Look at the sector breakdown. Consumption loans outstanding: Rs 1.26 trillion. Agriculture and forestry: Rs 354 billion. That's a 3.6-to-1 ratio of people buying things over people growing or making things. Wholesale and retail trade, essentially import financing, adds another Rs 1.04 trillion. Meanwhile, agriculture lending actually declined in the most recent review period, by 1.1%.

This is not a financing system aligned with how Nepal actually earns money. The country runs on remittances, over $7.5 billion came in during the first half of this fiscal year, up 32% in dollar terms. That cash sustains household consumption. Banks intermediated it into loans against that same consumption, and secured those loans mostly against land and buildings. NRB's own data puts the share of private sector credit backed by real estate collateral at 63.9%.

What you end up with is a system that bet heavily on property values staying liquid and household income staying stable. When either wobbles, the math gets uncomfortable fast.

The interest rate picture makes this stranger still. The weighted average deposit rate at commercial banks is 3.56%. Lending rates average 7.12%. By any standard measure, credit should be cheap and growing. In the first half of fiscal 2025/26, deposits grew 14.8% year-on-year. Private sector credit? 6.7%.

That gap is the story. Money is piling into the system and sitting there because banks don't see enough borrowers worth lending to, and borrowers don't see enough certainty to borrow. NRB has been absorbing surplus liquidity on a massive scale through its Standing Deposit Facility essentially paying banks to park money rather than push it out the door.

Low rates and abundant liquidity are the standard toolkit for stimulating credit. They're not working. That usually means the problem isn't the price of money. It's the confidence of the people who'd be using it.

September 2025 is part of the explanation, though probably not all of it.

The Gen Z-led protests triggered by a social media ban and anti-corruption anger turned violent. The state's response resulted in significant casualties. Property damage was formally assessed by a government committee at Rs 84.45 billion. What followed was the kind of political transition that makes business investment decisions very easy: you wait.

The Federation of Nepalese Chambers of Commerce and Industry said publicly that private sector confidence had reached an all-time low. The World Bank cut its Nepal growth forecast to 2.1% for FY26, citing the unrest and what follows from it. East Asia Forum's analysis, published in February 2026, noted that Nepal avoided immediate macro collapse, the external buffers held, but that the real damage was to micro-level confidence, the kind that determines whether a factory owner expands a production line or a developer breaks ground.

Macro indicators can look fine while the real economy hesitates. Nepal is doing exactly that. Inflation at 2.4%. Foreign reserves covering 18 months of imports. Current account in surplus. Balance of payments in surplus. All fine. And yet: no one is investing.

There's one more number worth sitting with. Non-banking assets, foreclosed collateral that banks have taken onto their own balance sheets after borrowers defaulted and auctions failed to clear, stood at Rs 51.3 billion as of mid-January 2026. A February report covering 19 commercial banks put their combined non-banking assets at Rs 46.6 billion, up nearly 20% year-on-year.

Banks are becoming landlords. Not by choice. When property markets are sluggish, collateral doesn't sell. When collateral doesn't sell, banks hold assets they can't easily value or dispose of. Those assets tie up balance sheet capacity, introduce uncertainty, and make risk officers nervous about the next loan. It's a loop, weak property markets make it harder to recover from bad loans, which makes banks more reluctant to lend, which slows economic activity, which keeps property markets weak.

Japan spent most of the 1990s in a version of this loop. So did parts of the eurozone after 2010. The details differ, but the pattern is recognizable: elevated NPLs left unresolved don't just sit there. They suppress credit and drag on output for years longer than they should.

None of this means Nepal is in a banking crisis. It isn't, not yet, not by any credible measure. Capital adequacy ratios are still above minimums, liquidity is genuinely abundant, and the external sector is providing real support. There's runway here.

But runway isn't a strategy.

What the situation calls for is fairly clear, even if doing it is hard. Banks need to treat NPL resolution as an operational problem, not a quarterly reporting one, with real workout teams, honest restructurings, and transparent disposal of foreclosed assets. The NRB's move toward ECL provisioning under NFRS 9 should push impairment recognition earlier in the cycle rather than later, which is healthy if banks integrate it into how they underwrite, not just how they report.

Credit allocation needs a different direction. Not by fiat, forced lending into agriculture has a poor track record, but by making it commercially attractive and practically viable to lend into productive sectors. That requires infrastructure, land tenure security, and value chain development that go well beyond what any central bank can deliver.

And the political side of this is unavoidable. Monetary policy can set rates and absorb liquidity, but it cannot manufacture the predictability that private investors need before they commit. Property rights, contract enforceability, policy continuity, these are what determine whether the next cycle of credit goes into things that generate income and jobs, or back into consumption and collateral.

Nepal's banking system has achieved real scale. The deposit base is large, the infrastructure is there, the liquidity is abundant. But a big system with weak loan quality and anxious borrowers isn't a sign of financial development. It's a sign of financial development stalling.

The next few years will show which direction it tips.