Bhutan's Banks: Half the Loan Book Is Buildings

Take everything Bhutan’s financial institutions have lent — every bank, every non-bank lender, every sector — and ask what it is secured on.

The answer, for close to half of it, is a building.

The housing share, and which way it is moving

At the end of March 2026, housing loans outstanding across all financial institutions stood at Nu 77.36 billion against total credit of Nu 276.89 billion — 27.94 percent of everything lent in the country.

That is the single largest category on the national book. It is also, and this is the part that gets missed, falling:

A slow, steady drift downward over fifteen months. The housing book itself is not shrinking — it is roughly flat at Nu 77 billion — but the rest of the credit system is growing around it. Whatever is being lent in Bhutan in 2026, it is not mainly mortgages.

That matters because the standard framing — housing is a third of the book and rising — is the one most commentary uses, and it is out of date in both level and direction.

What is replacing it

If housing is losing share, something is gaining it. The largest part of the answer is the category this series has already taken apart: service and tourism lending, which stood at roughly Nu 57.5 billion by June 2026 — about 21 percent of total credit.

So the two property-backed categories together:

Roughly half of all credit in Bhutan is secured on structures — homes on one side, hotels and commercial premises on the other.

At Bank of Bhutan, the largest lender, the concentration is sharper still. Its audited 2025 accounts show housing at Nu 29.80 billion and service and tourism at Nu 19.55 billion, out of a Nu 96.25 billion book. That is 51.3 percent — more than half of the country’s biggest bank, lent against buildings.

Why the two halves behave differently

This is not one exposure. It is two, and they are failing — or not — in different ways.

The hotel half has already broken. Tourism non-performing loans at Bank of Bhutan went from Nu 30.6 million to Nu 635.7 million across 2025 and reached Nu 1.149 billion by June 2026, against a national hotel stock running at around 20 percent occupancy. That is a construction boom that met no demand, and the losses are now arriving on the book.

The housing half is holding — but look at why. Housing NPLs at Bank of Bhutan rose 76 percent in 2025, from Nu 389.6 million to Nu 686.3 million, yet the default rate is still only 2.30 percent. The reason is not that housing borrowers are comfortable; it is that the ones who exist were selected to be. Only about 9 percent of Bhutanese households hold a housing loan at all. Banks lend 80 percent against value, so every borrower has already produced a 20 percent deposit — fifty-four months of median household income on a typical Thimphu flat. The housing book is small relative to the population precisely because the gate is so high, and the borrowers who got through it are the ones most able to pay.

Put differently: the housing half is performing because most of the country was never let in.

The concentration nobody prices

Here is what ties the two together, and it is the point of the whole series.

Half the national loan book is secured on property. And Bhutan has no observable property prices — the best land never trades, parcels under ten decimals cannot be registered, and declared transaction values are tax figures rather than market ones. The collateral behind roughly Nu 135 billion of lending has been valued by three parallel official systems and tested by almost no actual sales.

When it was tested — a Trongsa resort carried at Nu 280 million against Nu 179.5 million of bank exposure — it realised Nu 9.5 million, five percent of principal, after eleven years.

A banking system concentrated on an asset class is an ordinary thing; most small economies look like this. A banking system concentrated on an asset class whose value has never met a buyer is a different thing, and Bhutan does not appear to have asked itself which one it is.

A note on the data

The figures above are from the RMA’s March 2026 Core Indicators, which is the last complete print. The April and May 2026 releases exist but omit one institution entirely — Bhutan Development Bank’s row is blank — so their system totals cannot be used. That is worth knowing, because it means the most recent published picture of the national loan book is, as of late August 2026, five months old and the two releases since are unusable for exactly the question this piece asks.