Toam's Credit Book: How Fast Can It Grow Before Collateral Quality Slips?

The main article made the broad point: group growth in 2025 ran through Toam, but it also tightened the balance sheet. This follow-up isolates the question that point only hinted at. Not whether the credit book grew, but what kind of collateral now sits underneath that growth.

The picture is more complicated than the headline of a NIS 516.2 million credit book and only NIS 3.1 million of credit-loss expense. On the one hand, this is not a book that currently looks impaired: the cumulative allowance stands at NIS 10.0 million, revenue diversification is still reasonably broad, and Toam still has wide access to bank funding lines. On the other hand, the book clearly moved in a less conservative direction: real estate became heavier, the weight of first-lien collateral fell, and more of the money now sits with a relatively small number of large borrowers.

That is the key point. Toam did not just grow. It grew into a book whose quality depends more on appraisals, lien rank, and underwriting discipline, and less on the granularity of short-dated import-finance deals.

Growth came from the less granular part of the business

The first shift, and probably the most important one, is not the size of the book but its economic composition. Net credit exposure rose to NIS 516.2 million at the end of 2025, up 24.5% from 2024. But almost all of that jump came from real-estate finance, which rose to NIS 365 million from NIS 246 million, a 48% increase. Import finance actually fell to NIS 151 million from NIS 168 million.

This is not just a shift between two business lines. It is a shift between two risk profiles. Import finance is spread over 3,357 files, with tenors of 30 to 180 days and an average customer exposure of NIS 1.5 million. Real-estate finance is only 36 files, with an average customer exposure of NIS 10 million and tenors of one to five years. In other words, 2025 growth did not come from the shorter, more diversified, more operating-like part of the book. It came from the longer-duration, heavier, less granular part.

There is also an important gap between the income side and the balance-sheet side. In 2025, about 60% of financing-segment revenue came from real-estate borrowers and about 40% from commercial credit. In 2024 the split was 57% versus 43%. Put differently, the balance sheet became much more real-estate-heavy faster than the income statement did. The book took more risk before the revenue mix fully showed it.

New business flow points the same way. Real-estate credit extended during the year jumped to NIS 229 million from NIS 99 million, while annual import-finance volume fell to NIS 578 million from NIS 604 million. The 2025 growth engine was real estate, not import finance.

Collateral ranking moved backward, not forward

This table tells the quality story better than the growth rate does. At the end of 2024, first-lien credit stood at NIS 159.0 million, or 38.3% of the book. By the end of 2025 it had fallen to NIS 136.9 million, only 26.5% of the book. The more conservative layer shrank both in absolute terms and as a share of the total book.

By contrast, second-lien credit jumped to NIS 134.0 million from NIS 78.3 million, while second-lien lending on project surpluses rose to NIS 38.1 million from NIS 9.2 million. Together, those two layers already account for NIS 172.1 million, or 33.3% of the book, more than the entire first-lien layer.

The largest category is still third-party guarantees, floating charges, and other collateral at NIS 217.2 million, or 42.1% of the book. That does not automatically mean weak credit. But it does mean a large part of the book sits outside the simplest comfort zone of direct first-lien security.

This is where management’s own presentation creates the right tension. The capital-markets slide highlights only NIS 3.1 million of credit-loss expense in 2025, about 0.7% of the average credit book. At the same time, the collateral stack moved in a more aggressive direction. Both things can be true at once. Low losses in 2025 show the book has not yet been hit. They do not show that the new book was built on the same quality of protection as the old one.

The LTV picture makes the issue even sharper

In the real-estate book, the company itself publishes two LTV calculations. That is not a technical footnote. It is central to the read. Under option A, which takes into account limitations imposed by the senior lender on collateral value, NIS 113.1 million, 31% of the real-estate book, sits at LTV of 85.1% and above. Only NIS 34.6 million, 9.5% of the book, sits at up to 50% LTV.

That is already heavy. But what matters even more is that the risk focus does not disappear when moving to option B, management’s internal risk-assessment method. True, NIS 38.1 million of project-surplus financing shifts into the up-to-50% bucket there instead of staying in the high-LTV buckets. But NIS 98.4 million of second-lien lending still remains in the 85.1%-and-above range. Even under the friendlier internal lens, the more aggressive part of the book does not disappear. It mostly changes address inside the calculation.

The second issue is that LTV quality also depends on measurement quality. The company says it generally relies on appraisals, actual transactions, or zero reports, but when updated appraisals or reports are not available, collateral value is determined using management estimates based on objective market information. In a smaller book that may be easier to absorb. In a book where real-estate finance already exceeds NIS 364 million and where a large part of growth sits in second-lien structures, this is no longer a marginal footnote. It is part of credit quality.

The top-ten borrower table makes that even more concrete. The largest single exposure, NIS 51.25 million, or 9.9% of the book, is a real-estate borrower secured by a second lien at 96% LTV. Another real-estate borrower appears at 93% LTV with second-lien collateral, and another at 101% LTV with second-lien collateral. This is no longer an abstract discussion about averages. These are large borrowers whose protection already sits lower in the project capital stack.

Diversification looks wide until you check where the money sits

On paper, 136 borrowers sounds diversified. In practice, only 18 borrowers, 13.2% of the customer count, account for NIS 334.0 million, or 64.7% of the book. That is very high concentration in exposure terms even if the headline borrower count looks broader.

That gap between revenue concentration and balance concentration matters. The top 10 borrowers generate 37.9% of segment revenue, while the top 10 borrowers by exposure hold 45.1% of the book. The revenue line therefore looks more diversified than the actual risk line. Anyone reading only the revenue table can come away with a cleaner impression than the balance-sheet concentration really supports.

Real estate dominates here too. Nine of the 10 largest borrowers by outstanding balance are real-estate borrowers. The only commercial borrower in that top 10 relies on an owner guarantee, not on mortgaged real estate. In plain terms, the debate about book quality is really a debate about the quality of the real-estate book, not of the financing segment as a whole.

Why this matters now: growth was funded by banks almost one for one

The final point is that this debate is not academic. In the non-bank financing segment, net customer exposure grew by NIS 101.6 million in 2025. Over the same period, bank credit rose by NIS 105.5 million, while supplier credit actually fell by NIS 13.8 million. In other words, growth in the book was funded almost one for one through the banking system.

That connects directly to the funding structure. Toam has five uncommitted short-term bank facilities totaling about NIS 970 million, has granted a general floating charge in favor of its lenders, and is committed to a tangible-equity-to-tangible-assets ratio of at least 15%. At the group level there is also a 15% equity-to-assets covenant. The company says it is in compliance and also notes that it still has more than NIS 0.7 billion of unused bank lines. That provides room.

But it also sharpens the 2026 test. As long as the credit book keeps growing through real estate, through second-lien structures, and through a relatively small set of large borrowers, book quality is not only a question of eventual credit losses. It is also a question of funding cost, banks’ willingness to keep extending lines, and how much slippage the system is willing to tolerate before becoming more conservative.

Conclusion

In the main article, the broad story was growth through Toam alongside a tighter balance sheet. This follow-up compresses that into a sharper sentence: the weakness in Toam’s credit book is not currently visible through losses, but through the collateral quality sitting underneath growth.

This is still not a broken book. Current provisioning is low, the company is in covenant compliance, and bank access remains wide. But the quality of 2025 growth is lower than the headline of a bigger book suggests. Real estate now accounts for more than 70% of the book, first-lien weight has eroded, and exposure sits with a relatively small number of large borrowers.

From here, three checkpoints will determine whether the thesis stays manageable or gets heavier: whether first-lien weight stops shrinking, whether credit losses stay low even with a much more real-estate-heavy and concentrated book, and whether the next leg of growth arrives without a further move toward high-LTV and weaker collateral structures. If those conditions hold, 2025 will look like a controlled jump. If they do not, it will turn out that Toam grew faster than the cushion underneath it.