Leumi and Residential Real Estate: Where Credit Friction Starts

Where The Friction Starts

In the main article, this showed up as a sharp but compressed warning. This follow-up isolates that area on its own. Not the whole credit book, not the whole housing market, but the point where credit quality starts to grind before a broad credit-loss event appears: slower sales, longer project timelines, 20/80-style contracts, and a higher reliance on contractor-backed support to get deals closed.

What makes this interesting is that Leumi's yellow flag still does not look like a sudden credit event. Quite the opposite. The bank is still expanding exposure to the sector, says it remains within regulatory limits, and says the new risk-weight treatment for certain projects has no material effect on its capital ratios. That is exactly why the numbers need a closer read. When housing-market credit friction begins, it usually shows up first in sales velocity, deal structure, problematic-accruing credit, and capital consumption on loans that still look acceptable from the outside.

That is also the difference between a dramatic headline and an analytical read. If this were already a broad credit problem, the first signals would likely be a sharp jump in non-accrual loans, immediate capital pressure, or widespread deterioration in household mortgages. That is not the story yet. The story is that bigger parts of Leumi's housing and project-finance exposure now need more time, more indirect support, and a tighter regulatory frame to keep moving.

This Is Still Not A Mortgage Accident

The first thing to separate is three different pockets: construction, real-estate activity, and household mortgages. The numbers show that the sharper friction sits mainly in real-estate activity, not in the classic household mortgage book.

In real-estate activity, total credit exposure rose to NIS 61.488 billion from NIS 55.413 billion. Problematic credit in that pocket jumped to NIS 825 million from NIS 380 million. That is close to a doubling, and the problematic-credit ratio rose to about 1.34% from 0.69%. That is no longer a rounding issue. By contrast, in construction, problematic credit increased much less, to NIS 1.073 billion from NIS 1.023 billion, and as a share of the book it actually fell to about 0.68% from 0.77% because total exposure grew faster. Not every part of Leumi's real-estate exposure is behaving the same way.

Household mortgages still look more contained. Total housing-loan exposure rose to NIS 160.888 billion from NIS 150.383 billion. Problematic credit rose to NIS 789 million from NIS 700 million, and the ratio moved only to about 0.49% from 0.47%. That is deterioration, but on a very different scale. The amount of loans under payment deferral also fell to about NIS 2.4 billion at the end of December 2025, versus NIS 3.6 billion at the end of September 2025. That matters. The main friction point still does not look like broad household weakness.

This is the core point. Anyone looking for the first crack in the wrong place will miss it. Leumi is not signaling a broad mortgage problem at this stage. It is signaling that project economics and sales mechanics have become more crowded.

Once Contractors Need To Support The Sale, The Regulator Is Already In The Room

The bank describes the pressure chain directly: labor shortages, longer construction periods, higher execution costs, higher financing costs, more non-linear 20/80-style contracts, and slower sales. That is one sequence, not a random list. The harder a sale becomes, the more the contractor has to soften deal terms. The softer the deal terms become, the more carefully both the bank and the regulator start reading the credit behind them.

On April 6, 2025, the Bank of Israel limited contractor-subsidized bullet and balloon mortgages to 10% of quarterly housing-loan execution. Leumi says it was within that limit at year-end 2025. That is important, but it does not clear the issue. It only means the bank did not cross the regulatory red line. The temporary order itself shows that the regulator identified these structures as a point of friction worth containing.

There is also a second layer, and it is more interesting. Under the regulatory update, project-finance credit for residential construction gets a 150% risk weight if more than 25% of apartment-sale contracts defer more than 40% of the selling price until delivery. That is not a credit loss. It is a capital-cost increase for the same sale structure. So even if a project still looks alive in headline sales data, it is already consuming more capital. Leumi says the new rule has no material effect on its capital ratios. That is plausible for a bank of this size. But at the level of project economics, the message is clear: some deals are becoming more expensive to fund and more expensive to carry.

That is the distinction between activity that still looks strong and activity whose quality is already eroding. A contractor that sells apartments on 20/80 terms may preserve signing velocity. It does not necessarily preserve the same cash quality, execution certainty, or underwriting comfort for the lender.

Loan-To-Value Tells The Same Story

The loan-to-value and absorption-capacity table adds another layer. Within residential construction exposure, NIS 27.843 billion sits in the 65% to 80% loan-to-value bucket, and another NIS 1.825 billion sits above 80%. Together that is about 81% of the under-construction residential exposure. That does not mean the bank is operating without buffers. On the contrary, the report notes that most of the above-80% balance is also backed by additional collateral and or another repayment source. But it does mean that the margin for error depends less on home-price appreciation and more on project timing, sales velocity, and execution discipline.

That matters because 2025 gave developers exactly the less-forgiving combination: rising unsold inventory, almost flat apartment prices, price declines in some demand areas, persistent labor shortages, and financing costs that remained high versus the past. In that setting, even without a sharp home-price decline, a project becomes less forgiving to delays.

It is also worth noticing what is not disclosed. Leumi does not quantify its direct exposure to projects that cross the 25% deferred-payment threshold, and it does not provide a numerical breakdown of contractor-subsidized mortgage volume within total execution. So it would be wrong to claim that the bank is sitting on a large and clearly identified problem. But it is fair to say that the place where the problem would show up first is already clearly marked: projects moving through deferred-payment mechanics rather than through normal sales velocity.

What The Market Needs To Watch Now

The first test over the next 2 to 4 quarters is not whether Leumi stays profitable. That is almost taken for granted. The real test is whether problematic-accruing credit inside real-estate activity stabilizes at the friction layer, or begins migrating into non-accrual. That is the line between discomfort and genuine deterioration in credit quality.

The second test is the quality of any sales recovery. The risk report assumes rate cuts could support some recovery. If the recovery comes through real demand and inventory absorption, pressure should ease. If it comes mainly through more generous payment structures and contractor support, Leumi may continue lending, but the quality of the book will not really clear.

The third test is the relationship between exposure growth and protection growth. In 2025, problematic credit in Israeli construction and real estate rose by NIS 495 million, while the allowance balance rose by only NIS 238 million. That does not mean the bank is under-reserved. It does mean management is still reading the story more as prolonged uncertainty than as a realized loss event. If the trend continues, that gap becomes harder to leave open.

The fourth test is off-balance-sheet risk. More than half of the 2025 increase in total sector exposure came from other off-balance credit exposure, which rose by NIS 15.734 billion. That is a useful reminder that in real-estate lending, pressure does not start only once cash is fully on the balance sheet. It starts when there are more commitments, more project support, more guarantees, and more need to keep projects moving.

Counter-Thesis

The strongest objection to this harsher read is straightforward: maybe this is only a temporary adjustment to a market hit by war, high rates, and regulatory transition, not the beginning of a broader quality slide. The bank remains within its limits, says the new risk-weight rule is not material to capital, insures part of its sector exposure, and still shows fairly controlled household-mortgage data. The macro view in the report also points to lower rates over the next year, which could absorb a meaningful part of this friction without a broad credit event.

That is a good counter-thesis. It could also prove correct. But for it to hold, the issue has to remain in transition mode: friction in closing sales, not repayment failure; partial relief through lower rates, not another worsening in deal terms; and a localized rise in problematic credit, not a sequence of quarters where it keeps climbing specifically inside real-estate activity.

Conclusion

In the main article, this area appeared as a warning note. Read on its own, it is more than that. Leumi's residential real-estate credit friction does not start with default. It starts in a market that now needs more time, more contractor support, and more capital to produce the same sale. That is why the first move is showing up in problematic-accruing credit and in real-estate activity, not through a broad break in household mortgages.

This matters because a large bank does not crack in one day. First the economics of the deal weaken. Then capital consumption rises. Only later, if at all, does a sharper profit hit show up. Anyone trying to locate the first real crack in Leumi needs to look exactly there.