Under The Surface At BlackEdge: What The Jump In Stage 2 And Restructured Loans Really Means
BlackEdge ended 2025 with NIS 44.4 million in Stage 2, NIS 23.7 million of restructured loans and NIS 72.6 million of past-due balances. This follow-up tests whether that spike was a temporary transition inside a growing book, or an early sign that credit quality now depends on rollovers, time and recovery assumptions more than the headline provision suggests.
What This Follow-Up Isolates
The main article framed the core debate around BlackEdge: funding improved, but margin has not yet recovered. This continuation narrows the lens. The question here is what really happened to credit quality underneath the headline provision line.
The tables point to a less comfortable read than the top-line story. This is still not an open credit blow-up, but it is no longer statistical noise either. Stage 2, meaning exposures with a significant increase in credit risk, almost quintupled. Restructured loans moved from a marginal line item to a meaningful one. Delinquency aging also moved inward into the 31-day to 180-day buckets, not only into the old tail of balances already more than a year overdue. Put differently, 2025 looks less like a clean growth year and more like a year of active pressure management inside the book.
| Metric | 2024 | 2025 | Why it matters |
|---|---|---|---|
| Stage 2 balance | NIS 9.4 million | NIS 44.4 million | The jump far outpaced growth in the book itself |
| Restructured loans | NIS 1.3 million | NIS 23.7 million | A meaningful share of the exposure already needed revised terms |
| Loans with contractual due dates changed from the original schedule | NIS 3.17 million | NIS 29.4 million | The repayment schedule no longer looks as stable as it did a year earlier |
| Past-due balances | NIS 22.9 million | NIS 72.6 million | The pressure spread far beyond the old tail of problematic debt |
| Stage 3 coverage ratio | 30.16% | 20.1% | The next question is about recovery and realization assumptions, not only exposure size |
Stage 2 Jumped Much Faster Than The Book, And That Is The Warning Signal
BlackEdge's gross credit book grew 23% to NIS 970.3 million. That matters, but it is not the core point here. The core point is that the Stage 2 balance jumped to NIS 44.4 million from NIS 9.4 million, up 370%, while the Stage 3 balance rose to NIS 29.4 million from NIS 17.2 million, up 70.7%.
The implication is straightforward. The big move in 2025 was into the watchlist, not only into debt already defined as impaired. Stage 2 already represented 4.6% of the gross book, versus 1.2% a year earlier. If Stage 2 and Stage 3 are combined, the result is NIS 73.7 million, or 7.6% of the book, versus only NIS 26.6 million and 3.4% of the book in 2024. That is too sharp a shift to explain away as normal growth.
The less intuitive part is that this jump barely showed up in the Stage 2 allowance itself. The Stage 2 allowance ratio stood at 0.23% at the end of 2025, versus 0.31% a year earlier. Even the Stage 3 allowance ratio fell to 20.1% from 30.16%. That does not prove the allowance is too light. The company explicitly says its model relies on expected cash flows, collateral value and realizability, and on an ECL framework built around probability of default and loss given default. But it does mean that the 2026 debate will not be only about exposure volume. It will also be about the quality of the assumptions that still allow low coverage ratios even as balances move into weaker risk buckets.
Restructured Loans Are No Longer A Footnote
The most striking number under the surface is not necessarily Stage 3. It is the amount of rolled or reworked debt. In the expected credit loss table, restructured loans rose to NIS 23.684 million from only NIS 1.272 million. That means more than half of Stage 2 at the end of 2025, about 53.4%, already sat on debt that had gone through revised terms.
At the same time, the maturity table shows a broader disclosure: loans whose contractual due date changed from the original schedule reached NIS 29.408 million, versus only NIS 3.17 million at the end of 2024. The two figures are not identical, but they point in the same direction. The company is not only identifying higher risk. It is also giving more loans more time.
What makes this more revealing is that the average life of the whole portfolio barely moved, to 0.79 years from 0.78 years. Anyone looking only at the average maturity could miss the story. The issue is not sitting in the macro number for the entire portfolio. It is sitting in specific pockets that now need delay, rescheduling and tighter monitoring.
That is also the difference between a clean transition year and an early warning year. A transition year can include isolated restructurings. It should not usually include jumps of tens of millions of shekels in two parallel disclosures. When both the Stage 2 table and the changed-due-date table move in the same direction, it becomes harder to call this operating noise.
Delinquency Aging Shows The Pressure Moved Deeper Into The Book
The jump in monitored exposures did not remain theoretical. Past-due balances rose to NIS 72.635 million from NIS 22.881 million, up 217%. But the more important point is how that pressure moved across the aging buckets.
The 0 to 30 day line actually fell to NIS 271 thousand from NIS 1.331 million. That detail matters because the company itself says this bucket also includes four clearing days, so it can be noisy. The heavier buckets, by contrast, moved sharply: balances 31 to 60 days past due rose to NIS 12.255 million from NIS 519 thousand, balances 61 to 120 days past due rose to NIS 23.462 million from NIS 3.227 million, and balances 121 to 180 days past due rose to NIS 9.801 million from NIS 1.239 million. The long tail also grew, with NIS 11.688 million already more than 365 days past due versus NIS 7.192 million a year earlier.
The point is not only that delinquency increased. The point is that the pressure shifted away from the shortest and noisier buckets, where technical timing and clearing can distort the picture, and into ranges that say more about actual repayment strain. That is why 2025 is hard to read as a year where the problem was already resolved through collateral and rollovers.
The Default Policy Explains Why Stage 2 Is The Key Metric Right Now
The company's accounting policy sets a relatively demanding threshold for default. For a balance to be treated as impaired, two cumulative conditions are required: contractual payments must be more than 120 days past due, and there must also be another adverse event affecting expected cash collection, such as significant financial difficulty, a breach of contract, a concession granted because of financial stress, or an expected insolvency or financial restructuring process.
That matters a great deal for how 2025 should be read. If the threshold for entering Stage 3 is relatively high, Stage 2 becomes the place where pressure accumulates before accounting default is fully recognized. That is why the jump in Stage 2 and in restructured loans should not be dismissed as a technical detail. This is precisely the zone where the company is still trying to stabilize, roll and collect before the issue turns into explicit credit impairment.
The same logic also explains the tail above one year. The company says it writes off financial assets that are more than one year past due only when no further collection actions are being taken, or when there is no reasonable expectation of recovery. So the mere existence of NIS 11.688 million already more than 365 days overdue does not contradict the stated policy. But it does mean that the credit book still carries an older layer of debt that remains alive only because recovery or realization is still assumed.
That is exactly where the difference sits between an overly relaxed read and a careful one. If recovery and collection work, 2025 may indeed look back like a transition year with a lot of manual management. If they do not, part of what is currently sitting in Stage 2 or deep delinquency will have to move one step further in accounting recognition.
What Has To Happen In 2026 For This To Still Look Transitional
The less comfortable news is that the pressure did not fade by year-end. Out of the NIS 11.215 million annual credit-loss expense, about NIS 4.96 million was recorded in the fourth quarter alone, roughly 44% of the full-year total. In other words, 2025 did not end with a clear easing of pressure. It ended with a relative acceleration into year-end.
That also defines the right checkpoints for 2026. It will not be enough to show continued book growth or lower funding costs. The company will have to show that Stage 2 actually comes down, that restructured loans return to a normal repayment path rather than simply staying rolled over, that the 31-day to 180-day delinquency buckets shrink, and that the more-than-one-year tail stops growing. Without that, any margin improvement will still be only part of the story, because the credit book will continue to carry a friction layer waiting to be resolved.
Bottom Line
The jump in Stage 2 and in restructured loans means BlackEdge ended 2025 with a credit book that requires more management, more time and more trust in recovery assumptions than the headline provision number alone suggests. This is still not proof of a credit crisis. But it is clear evidence that the 2026 debate cannot remain only about funding cost and margin.
If 2025 was a true transition year, the coming reports will need to show lower monitored exposure, fewer accumulated rollovers and improving deep-delinquency buckets. If that does not happen, the right way to read the Stage 2 jump will no longer be as a one-off transition effect, but as an early signal that the real issue sits in rollover quality, not only in the price of funding.
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.