Cambodia Housing Credit and Real Estate Indicators

“Real-estate credit” is not a single self-explanatory number in Cambodia’s official statistics. In Table 15 of National Bank of Cambodia Economic and Monetary Statistics Bulletin No. 385, `Of which Mortgages, Owner-Occupied Housing only` appears directly beneath `Real Estate Activities`. For February 2026, the broader real-estate line is KHR 52,968.5 billion and the mortgage line is KHR 23,635.0 billion. In that table, the mortgage figure is a subset, not a second amount that can be added on top.

That small label changes the interpretation of any headline about credit growth. Before comparing percentages, the reader needs the publication, reporting universe, reference period, unit and the relationship between categories. A number can be perfectly accurate and still support the wrong market story if one of those pieces is missed.

There is a second source of confusion. “Lending” is often read as money newly advanced during a month, but this table reports a gross loan balance at a point in time. It describes credit still outstanding on balance sheets. It does not tell us how many new mortgage contracts were signed, how much gross new credit was disbursed during the month or how many homes changed hands.

The publication month and the data month also need to be kept separate. Bulletin No. 385 is a March 2026 issue, but its sector-credit table reports balances through February 2026. Calling those figures “March credit” would shift the reference period by one month and can create a false comparison with a March price or sales series.

`Construction` is the credit line classified to construction within NBC’s sector framework. The February 2026 balance is KHR 24,136.9 billion. For a property buyer, that is best read as banking exposure to the construction sector at the reporting date. It does not isolate residential development, identify which projects received funding or show how much cash was drawn and spent on construction during that month.

The same table reports KHR 52,968.5 billion for `Real Estate Activities`, followed immediately by KHR 23,635.0 billion for `Of which Mortgages, Owner-Occupied Housing only`. The wording matters: the mortgage balance sits inside the broader parent line in this monthly presentation. Adding the two produces double counting. Subtracting the mortgage line from the parent gives roughly KHR 29,333.5 billion, but that is a derived residual, not a separately published NBC series, and it should not be assumed to match a bucket used elsewhere.

The 2025 Financial Stability Review illustrates why labels cannot be carried from one NBC publication to another without checking the presentation. Figure 2.6 displays Real Estate Activities, Construction and Mortgages as separate analytical sectors, at 12.4%, 10.4% and 10.0% of credit respectively. Figure 2.7 gives separate 2025 growth rates of 13.4%, 11.3% and -5.4%. Those charts are useful for annual composition and direction, but they do not rewrite the monthly table’s explicit `of which` relationship. The safe rule is to read each figure and table on its own terms.

Outstanding credit and new lending also answer different questions. Consider an illustrative portfolio that starts the month at 100. Banks make 10 of new loans while borrowers repay 15, leaving 95 outstanding at month-end. The balance has fallen by 5 even though gross new lending was 10. Real data can also move because of write-offs, reclassifications, valuation or foreign-exchange effects and revisions, so a month-to-month change in stock cannot be relabelled as gross new lending.

The public sources used here do not provide a clearly defined national time series of newly issued mortgages that can be placed beside the outstanding balance. The IMF’s 2024 technical assistance report says NBC uses residential-loan data to compile a report on newly issued property loans for its Financial Stability unit. That shows the underlying information exists inside the statistical system, but it does not establish a public series with a stable definition and release history. The appropriate result for this page is therefore to leave a new-lending row out instead of manufacturing one from changes in outstanding credit.

Reporting coverage is another place where apparently comparable numbers can diverge. Bulletin No. 385 labels Table 15 as credit granted by depository institutions and notes commercial banks and microfinance deposit-taking institutions; the following table covers non-depository institutions separately. A numerator drawn from one institutional universe and a total drawn from another can still produce a plausible-looking percentage, but it is not the same statistic. The reporting population belongs beside the date and unit whenever a share is interpreted.

Revisions deserve similar care. Bulletin No. 385 explicitly notes revisions to several tables for January and February. That note does not mean every credit cell has necessarily been revised, but it shows why a stored value should travel with its release and reference period. Two downloads of a series at different dates can differ because of revisions or reclassification even when the row label has not changed.

The rows below come from Table 15 of NBC Bulletin No. 385 and refer to February 2026; the unit is KHR billion. They are comparable within the same period and depository-institution universe, and the mortgage `of which` line must not be added to its `Real Estate Activities` parent.

Price evidence comes from a different series. NBC’s Financial Stability Review 2025 reports that the nationwide Residential Property Price Index was down 3.8% year on year in December 2025, while Phnom Penh was down 4.3%. In the same year, the review reports Real Estate Activities credit up 13.4%, Construction up 11.3% and Mortgages down 5.4%. Putting those directions beside each other is informative; turning them into a causal chain is not. More real-estate or construction credit does not prove that home prices rose, and a smaller mortgage balance does not prove that a particular apartment became cheaper.

RPPI itself has a defined coverage. The IMF’s 2024 technical assistance report explains that NBC’s index was built from residential-property loan records supplied by mortgage lending institutions. Cash purchases, developer-financed transactions and purchases funded by foreign banks sit outside that loan-based sample. The index therefore represents a loan-observed segment of the market, not a census of every residential transaction in Cambodia.

Condominiums are an especially important limitation. The IMF found too few condominium and apartment observations in the loan data to estimate price change reliably, so the published RPPI at the time of the methodology review covered houses. That distinction matters for many foreign buyers in Phnom Penh. A fall in the national or Phnom Penh RPPI should not be presented as a measured fall in Cambodian condo prices unless a newer official methodology explicitly expands that coverage.

Geography can create another false match. National credit is a country-wide banking measure; a Phnom Penh RPPI series is a capital-city price measure. They can be shown as separate pieces of context for the same period, provided the different universes are stated. They should not be forced into an equation such as “credit changed by X, therefore prices should change by Y,” and a short correlation would not establish which variable drives the other.

NBC also maintains RPPI as a separately updated statistical series, which is a useful reminder that the price index and the sector-credit table are different products. A more recent observation does not erase a coverage limitation by itself. Unless a newer official methodology explicitly shows that condominiums have been incorporated into the published index, the limitation documented by the IMF should remain visible.

The 2025 pattern is useful precisely because the signals do not line up neatly. Mortgage credit contracted, construction and real-estate credit expanded, and RPPI declined. That combination does not identify a driver, but it does demonstrate why borrower finance, developer or sector finance, and measured residential prices cannot be collapsed into a single “property market” indicator.

For a property buyer, these figures are useful at three different levels. At sector level, they show where deposit-taking institutions carry exposure and whether broad categories such as construction, real-estate activities and mortgages are expanding or contracting. That is useful cycle and risk context, especially when the categories are moving in different directions.

At market level, the data help prevent a single credit percentage from becoming a story about sales or prices. A shrinking mortgage balance alongside rising real-estate credit is possible because the borrowers and financing channels are different. A falling RPPI adds another signal, but one drawn from a different dataset with its own geographic and property-type coverage. The comparison becomes useful only after those boundaries are kept visible.

A decision on one apartment sits at a much narrower level. Sector credit cannot establish the fair value of the unit, the number of completed sales in its building, the developer’s solvency, the security granted to a lender, the project’s drawdown schedule or the buyer’s contractual priority. Those questions require property-level evidence: the contract, title and encumbrances, the entity receiving payments, construction progress, comparable units and, where available, completed transactions.

NBC credit statistics are therefore best used as a map of the financial backdrop, not as a buy or sell signal. They can tell a buyer which questions deserve more attention and which market narrative is too simple. Once the decision is about a named property, the evidence should move to the same level of detail as the decision itself.

This distinction is particularly useful when reading news stories and sales material. “Banks increased property financing” may be correct for one credit category while saying nothing about household mortgage demand. “Mortgages fell” may describe a smaller outstanding balance after repayments without revealing gross approvals or disbursements. An RPPI citation may be legitimate house-price context yet weak evidence for a condominium. The official number becomes decision-useful only when its boundary is carried into the conclusion.