Opal Balance: Is the New Mortgage Book Really Matched to Its Funding
At the end of 2025, Opal's mortgage book looked nominally covered: the company had NIS 110 million of committed dedicated facilities and another NIS 20 million of non-committed capacity against a net mortgage book of about NIS 94 million. But that is still a better-than-basic funding envelope, not yet a full duration and rate match.
What This Follow-Up Is Testing
The main article argued that Opal's core test has shifted from book size to funding quality. This continuation isolates the place where that argument matters most: the mortgage book. By the end of 2025, mortgages were already about 16% of the group's credit book, but only 8.5% of revenue. In other words, this is already a meaningful balance-sheet leg before it is a fully meaningful earnings leg.
The first read looks comfortable. The net mortgage book stood at about NIS 94.2 million, 91% of it sat in maturities of 1 to 4 years, and the company already had NIS 110 million of committed dedicated facilities for real-estate-backed lending, plus another NIS 20 million of longer-term non-committed capacity. On the surface, that looks like a book that is funded.
But that is only half true. The new book is not exposed like a long asset financed by an overdraft, but it is not matched one-to-one either. Opal's answer is a group-level funding envelope: equity, fixed-rate bonds, and bank facilities that reprice and renew faster than the assets. So the right conclusion at year-end 2025 is not that the book is unfunded. It is that the hedge looks numerically solid, but still short of a full duration and rate match.
- Two committed mortgage-specific facilities totaling NIS 110 million already covered the year-end net mortgage book.
- Including the NIS 20 million non-committed line, the dedicated mortgage envelope rises to NIS 130 million, or about 1.38x the net book.
- Yet 85% of the mortgage book carries fixed or fixed-linked pricing, while key bank funding lines reprice or renew over time.
- About NIS 62 million of the NIS 94 million mortgage book came from DOI, so most of the matching question is being tested on an acquired platform rather than only on a gradually built legacy book.
On Paper, the Coverage Exists
The strongest point in Opal's favor is that the mortgage book did not outrun its dedicated funding envelope. In the operating discussion, the company says the mortgage business is funded through a mix of equity, fixed-rate non-linked bonds, and bank facilities, and it refers to roughly NIS 130 million of facilities supporting the activity. Note 11 provides the breakdown: a NIS 60 million committed facility from Bank A for real-estate-backed loans, a NIS 50 million committed facility from Bank B for similar lending with an average period of up to 24 months, and a NIS 20 million non-committed long-term line from Bank D for up to 5 years.
That matters. At the end of 2025, Opal did not lack nominal funding capacity for the mortgage book. Even without counting the up-to-NIS 75 million facility that Opal committed to make available to DOI as part of the acquisition, the two committed facilities alone already exceeded the year-end net mortgage book. Add the non-committed line and there is visible headroom.
But nominal coverage is not the same thing as a clean economic hedge. First, NIS 20 million of the NIS 130 million envelope is non-committed. Second, Bank B's line is structured around an average term of up to 24 months, while Bank A requires each underlying loan not to exceed 5 years from origination. That is a reasonable operating fit for a 1 to 4 year book, but it does not lock the cost of funding for the full life of each asset. Third, the NIS 75 million DOI facility is an internal group commitment to the acquired subsidiary. It improves liquidity inside the acquired platform, but it is not a new external source of funds at the group level.
What Is Actually Built Correctly
| Layer | What the filings show | Why it supports the case that coverage exists |
|---|---|---|
| Dedicated committed facilities | NIS 60 million from Bank A and NIS 50 million from Bank B | The two committed lines already exceed the year-end net mortgage book |
| Additional backup layer | NIS 20 million from Bank D, long-term and non-committed | Adds flexibility, even if it is not equivalent to hard committed capacity |
| Fixed-rate funding at group level | Series D bonds carry a stated coupon of 5.5% | Adds a less prime-sensitive funding layer, even if it is not formally allocated only to mortgages |
| Funding-linked underwriting discipline | Bank covenants cap LTV at 70% in Bank A and 65% in Bank B | The lenders themselves imposed an underwriting frame that limits collateral slippage |
The banks also shaped the book directly. Bank A requires LTV of up to 70%, a credit-to-mortgage-activity ratio capped at 0.8, and a 5-year ceiling on each loan. Bank B caps lending at 65% of collateral value and NIS 5 million per borrower. These are not just collateral clauses. They show that the funding itself has been built around a relatively conservative asset profile.
Where the Match Still Breaks
The weaker point sits not in the size of the lines, but in the speed at which they can reprice relative to the assets. Note 16 is explicit here. The company says it is exposed to a gap between the term of the funding facilities used for mortgage activity and the longer term of the loans extended to customers. It also states that the bank funding serving the activity renews annually, and that if the lines are not renewed, or if they become more expensive beyond ordinary prime updates, the result can be a cash gap and pressure on profitability and equity.
That is the key sentence in the whole exercise. The mortgage book is not built on a one-to-one asset-liability lock. It is built on the assumption that bank facilities will remain renewable and reasonably priced. So this is not a story of a severe mismatch, but it is not a story of a fully locked hedge either.
The Rate Side Is Not Cleanly Hedged
Another source of mismatch is the rate mix. According to the operating section, about 53% of the mortgage book carries fixed non-linked rates, about 32% carries fixed CPI-linked rates, and only about 15% carries variable rates. That means 85% of the book does not reprice in step with prime.
On the other side, the company itself explains that the bank funding used for the activity reprices over time, and parts of it are linked to Bank of Israel rates or prime. There is some natural offset here. Series D bonds are fixed-rate, and management says it reviews the fit between sources and uses on an ongoing basis. But the company does not disclose a formal allocation of the bond proceeds specifically to the mortgage book, so it would be too strong to claim that the fixed-rate mortgage book is largely funded with fixed-rate liabilities.
The implication is straightforward. Opal hedged the mortgage book better than a lender relying only on short rolling lines, but it still has not locked in the financial spread on most of the book. If rates fall, that can work in its favor. If line pricing moves up, a large part of the asset side cannot adjust at the same speed.
The Liquidity Note Still Tells a Conservative Story
The liquidity table in Note 16 presents the NIS 269.1 million of bank credit under the up-to-half-year bucket. That table is not mortgage-specific, but it is still a useful reminder of the system Opal actually operates in: a system of rolling bank funding, not one of long-duration liabilities locked against every layer of assets. So even if the mortgage book looks covered at the level of the dedicated envelope, it still sits inside a group whose main funding base behaves like shorter funding.
What Still Protects the Asset Side
This is where the analysis should stay balanced. A partial funding mismatch does not mean the book is equally risky on the asset side. In fact, Opal enters 2026 with a few real cushions.
First, the duration is not as extreme as the word mortgages might suggest. The company says the average period of the mortgage book is about 3.2 years, 91% of balances sit in loans with maturities of 1 to 4 years, and only about 9% are expected to repay after 4 years. In other words, this is longer than the check-discounting book, but it is not a classic 20-to-25-year mortgage portfolio.
Second, the collateral side looks relatively conservative. The LTV table in Note 16 is presented on a contractual-balance basis that includes future interest components, which is why it is slightly above the balance-sheet carrying amount. On that basis, about 76.5% of the mortgage book sits below 60% LTV, and almost 98% sits below 70%. There is no exposure above 80% LTV in that table.
That does not eliminate funding risk, but it does explain why the banks are willing to provide the lines in the first place. The book is built with a reasonable collateral buffer and a medium, not extreme, duration profile. That is why the right conclusion is not that the book is unhedged. It is that the book is partially hedged and still relies on renewal and sensible pricing.
DOI Adds Another Layer to the Test
The final point is where the book came from. The investor presentation says roughly NIS 62 million of the NIS 94 million mortgage book is attributable to DOI. That means about two thirds of the book entered Opal through an acquisition, not through slow organic build-out. That makes the matching question more immediate because the acquired platform does not get a long grace period to grow into its funding structure.
This connects directly to the transaction structure. As part of the acquisition, Opal committed to provide up to NIS 75 million of financing to DOI, and it also recognized NIS 759 thousand of deferred consideration and NIS 3.328 million of Put-option liability on the remaining 30%. These amounts are not large enough to threaten the group, but they do matter conceptually. The new mortgage leg is not free. It arrives with its own layer of funding and capital commitments.
Conclusion
Opal's new mortgage book does look funded on paper, but not in the cleanest possible sense. By the end of 2025 the company already had NIS 110 million of committed mortgage-specific facilities against a net book of about NIS 94 million, and the extra NIS 20 million non-committed line provided visible nominal cushion. On top of that, the asset side looks relatively conservative: average duration of about 3.2 years, 91% of the book in the 1 to 4 year range, and a large majority of exposure below 60% LTV.
But this is still not a full duration and rate hedge. Most of the book is not variable-rate, some of the lines renew or reprice along the way, and the company itself states explicitly that there is a risk gap between funding term and asset term. So the more precise reading is this: in 2025 Opal largely solved the availability question for the mortgage book, but it did not fully solve the stability and pricing question.
If 2026 shows orderly renewal of the lines, stable funding spread, and a larger profit contribution from mortgages without a meaningful increase in provisions, the market will be able to read this as a properly funded second engine. If funding costs rise, if bank lines become less accommodating, or if the book grows faster than the dedicated envelope, then the claim that the book is matched will turn out to be accurate only at the level of nominal coverage, not at the level of full spread economics.
Disclosure: Deep TASE analyses are general informational, research, and commentary content only. They do not constitute investment advice, investment marketing, a recommendation, or an offer to buy, sell, or hold any security, and are not tailored to any reader's personal circumstances.
The author, site owner, or related parties may hold, buy, sell, or otherwise trade securities or financial instruments related to the companies discussed, before or after publication, without prior notice and without any obligation to update the analysis. Publication of an analysis should not be read as a statement that any position does or does not exist.
The analysis may contain errors, omissions, or information that changes after publication. Readers should review official filings and primary sources before making decisions.