PaymenT: Credit quality and the full arrears picture
PaymenT's default rate is still running at about 1%, but that is only the headline. At year-end 2025 another 1.9% of the aged book sat in 1 to 90 day arrears, so the real credit read has to combine the arrears buckets, the fair-value accounting treatment and the CPI-linked loan structure.
What Is Left After the Main Article
The main article argued that PaymenT is growing quickly, but the shift to a lighter-capital model is still incomplete. This follow-up isolates the question the whole story rests on: how clean the credit book really is. If underwriting is holding and arrears stay contained, the balance sheet can work harder without breaking the thesis. If the headline of roughly 1% default hides a wider arrears funnel, then fair value, portfolio assignments and growth quality all need to be read more carefully.
The answer upfront: the positive headline still holds, but it is incomplete. At the end of 2025, defaulted loans stood at NIS 1.48 million, about 0.8% of the book shown in the aging table. That is low. But there were also another NIS 3.59 million in 1 to 90 day arrears, so the fuller picture of loans that were not fully current was about NIS 5.1 million, or roughly 2.6% of that aged book. That does not break the high-quality credit thesis, but it does mean default alone is not enough.
There is a second layer that matters from the start. Most of the loans PaymenT holds are intended for realization and are measured at fair value through profit and loss. That means expected credit losses do not sit only in the standalone credit-loss allowance line. They are already embedded in the valuation of the loans. Anyone reading only the allowance line in the income statement can end up understating the real credit discussion.
The Full Arrears Picture
The most important number here is not the annual default rate but the year-end aging table. It shows NIS 186.4 million that was not overdue, NIS 3.59 million in 1 to 90 day arrears, and NIS 1.48 million in default. In other words, 97.4% of the aged book was current, 1.9% sat in early-stage arrears, and 0.8% had already crossed into default.
| Layer | Balance | Share of the book | What it means in practice |
|---|---|---|---|
| Not overdue | NIS 186.4 million | 97.4% | The core support for the argument that underwriting is still holding at larger scale |
| 1 to 90 day arrears, not in default | NIS 3.59 million | 1.9% | The early warning layer before loans move into default |
| Defaulted loans | NIS 1.48 million | 0.8% | The portion that already crossed the company’s default threshold |
What matters most is the definition, not only the number. Under the company’s accounting policy, a significant increase in credit risk begins when contractual payments are more than 30 days past due, unless there is reasonable evidence to the contrary. Default is stricter than that. A financial asset is considered in default only when two conditions are met together: payments are more than 90 days past due, and there is another negative event that harms expected future cash flows. That is a tighter hurdle than the simple question of whether the borrower is late.
This explains why the headline of about 1% default looks cleaner than the wider picture. The default metric is indeed low, but it is already after a fairly strict filter. Anyone trying to assess real credit quality has to watch the 1 to 90 day bucket as well. That is where the next few quarters will either confirm or weaken the thesis.
There is another important point: even defaulted loans are not yet the same as final loss. The company writes off financial assets only when they have been overdue for more than a year and no further collection actions are being taken, or when it believes collection odds are exhausted. That makes the 0.8% default figure relatively conservative, but it also means the path from early arrears to accounting loss can take time.
How To Read The Default Rate Without Getting It Wrong
The company’s transaction table shows a credit default rate of about 1% in both 2025 and 2024, and below 1% in 2023. It also says this metric is externally validated and audited by the financial institutions that buy its loans. That matters, because it means the figure is not just an internal model output.
But the accounting layer changes the read. In the 2025 results discussion, the change in the allowance for credit losses was only NIS 783 thousand, and management describes it as not materially changed. On a first read, that can make credit risk look almost frictionless. Yet the company also states explicitly that expected credit losses on loans held for sale are included in the fair value measurement of those loans. So the standalone allowance line tells only a small part of the story.
That gap becomes clearer on the balance sheet. As of December 31, 2025, loans intended for realization stood at NIS 186.7 million and were measured at fair value. By contrast, the receivable book measured at amortized cost was much smaller: NIS 5.73 million gross, with a NIS 934 thousand allowance and a NIS 4.8 million net balance. Put differently, most of the credit book sits in the world of fair value and expected sale, not in the world where every credit issue flows visibly through a classic allowance line.
That is why PaymenT’s credit quality has to be read across three screens at once:
| Reading layer | What it shows | Why it matters |
|---|---|---|
| Annual default rate | About 1% in 2025 | The headline underwriting read still looks stable |
| Year-end arrears buckets | 1.9% in early arrears and another 0.8% in default | The wider friction picture is larger than the headline |
| Fair-value accounting | Expected losses are already embedded in the valuation of loans held for realization | Credit quality cannot be read from the allowance line alone |
What The Underwriting Read Actually Says
The transaction data gives a fairly strong clue about the shape of growth. In 2025, originated volume rose to NIS 535 million from NIS 389 million in 2024, and transaction count jumped to 23,498 from 14,052. At the same time, the average ticket fell to NIS 22.8 thousand from NIS 27.7 thousand, and average duration moved down to 14 months from 15 months. Average nominal and effective rates also declined modestly.
The reasonable inference is that the company did not buy growth through larger ticket sizes or longer duration. It did the opposite. Growth came through more transactions, with smaller average size and slightly shorter duration. That is not proof on its own of conservative underwriting, but it is consistent with a book that is becoming broader and more granular rather than more stretched.
There is a counterpoint. The share of credit sold to purchasing institutions fell to 60% in 2025 from 90% in 2024. So even if underwriting quality did not deteriorate, a larger portion of the credit stayed closer to PaymenT’s own balance sheet for longer. That is exactly where credit quality moves from being merely reassuring to being critical. The company did note that as of the report date, the loans intended for realization were around NIS 187 million, and roughly NIS 80 million of that amount was expected to be assigned during the first quarter of 2026. But until those assignments close in practice, the risk still sits with the company.
One more small but meaningful detail sits inside the risk-management description: in some transactions, merchant payout is released in stages and according to defined criteria. That means PaymenT’s underwriting does not stop at approval. It continues through the way proceeds are released to the merchant. This helps explain how the company may be keeping defaults low even though the book is mostly consumer credit and is largely unsecured.
CPI Linkage Is Both A Tailwind And A Risk
This is the layer many readers can miss. Most of the loans PaymenT originates carry a fixed interest rate and are linked to the CPI. By contrast, its bank funding is floating-rate, based on prime, and not CPI-linked. That creates a two-sided structure: higher CPI lifts interest and linkage income for the company, but it can also raise the borrower’s monthly installment and therefore increase credit risk.
The company quantifies both sides. Its sensitivity analysis shows that a 1% increase in the CPI adds about NIS 1.816 million to the fair value of loans intended for realization. On the other hand, a 1% increase in interest rates affects short-term bank credit in the opposite direction by about NIS 1.042 million. In plain terms, the loan book can benefit from inflation, but the funding base and the consumer-risk side can push the other way.
Management argues that as inflation has moderated, the risk of rising credit losses has also fallen, and that it keeps repricing credit to preserve spreads. That is a reasonable position, but it also identifies the real risk checkpoint: a low default rate today does not eliminate the fact that the borrower is paying a CPI-linked installment. If inflation surprises to the upside again, the first place to watch will not be revenue but the 1 to 90 day arrears bucket.
That said, it is important not to overstate the issue. The company also says its overall exposure to CPI movements is not material at the total market-risk level. So this is not a thesis about a hidden CPI bomb on the balance sheet. It is a thesis about a credit system where CPI linkage improves part of the business economics as long as the borrower keeps paying.
What The Market Should Measure From Here
From here, three checkpoints matter. First, whether the 1 to 90 day arrears bucket remains a manageable noise layer or starts feeding into default. Second, whether PaymenT continues taking loans off balance sheet at a pace that matches origination, especially after a year in which the share sold to purchasing institutions fell. Third, whether a more benign inflation and rate backdrop keeps helping, or whether borrower repayment capacity starts to show stress again.
Market data does not currently signal an aggressive public-market challenge to this point. In the latest short-interest snapshot, short float stood at 0.00% and SIR at 0, versus sector averages of 0.51% and 1.157. That means there is no meaningful short bet against the company’s credit story right now. But the absence of short interest is not the same as outside validation. It simply means the real argument will keep running through the filings.
Conclusion
The thesis remains constructive, but it needs to be framed more precisely. PaymenT is still showing good credit quality: only about 0.8% of the aged book was in default, and the annual default rate remained around 1% in a year of strong growth. But the full picture is wider, because another roughly 1.9% sat in 1 to 90 day arrears, and most expected credit loss is running through fair value rather than only through the allowance line.
So the right question into 2026 is not whether PaymenT can report a low default rate. It already has. The real question is whether it can keep that same cleanliness while growing, while temporarily carrying more credit on balance sheet, and while continuing to originate mostly CPI-linked loans. If the answer remains yes, the high-quality credit thesis strengthens. If the early arrears layer starts to widen, that will be the first place where the story begins to change.
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.