Amidar: The Rating Looks Sovereign, but Where Exactly Is the Line Between State Support and a Guarantee

What This Follow-up Is Isolating

The main article argued that the government backstop is what stabilizes Amidar's credit story. This follow-up isolates the sharper question underneath that claim: is that backstop a state guarantee, or something narrower. That is not legal trivia. It determines whether Series A rests on a direct undertaking by the state to the bondholders, or on a payment mechanism that runs through Amidar alongside a very strong assumption that the state will not let the company fail.

The local evidence is unusually consistent on this point. Midroog's follow-up report says the rating is determined under its GRI, Issuer Related Government, approach and without setting a standalone BCA for Amidar. The annual bondholder disclosure says explicitly that there is no guarantee for the company's obligations under the trust deed, and also that the bonds are not secured by collateral. Then note 18(c)(1) adds the sharpest legal line of all: the debt-raising agreement is between the government and Amidar only, and does not create rights for any third party, including the bondholders.

That is where the correct reading starts. The rating looks sovereign, but the legal claim is not sovereign. What exists here is a very strong state-support structure, an orderly payment mechanism, and evidence that the state has already transferred money through that mechanism. What does not exist is a direct commitment by the state to the bondholders as if the state itself were the obligor to them.

Three Layers That Should Not Be Blended

The easiest way to misread Amidar's bonds is to blend three different layers into one:

That is the core distinction. In credit markets it is very easy to see Aaa.il and stop there. Here the right order is the opposite: understand the structure first, then understand why the rating reaches that level.

What The Agreement Actually Gives Bondholders

Under the debt-raising agreement, the state committed to transfer the bond payments to Amidar according to the amortization schedule. The money is supposed to reach the state-payments account, which is owned by the company, no later than 14 business days before each payment. The company also undertook not to pledge those state-payment funds and not to use them for anything other than payments to bondholders.

There is also a reserved fund. Note 18(c)(1) and the attached valuation work explain that back in June 2018 the state transferred NIS 50 million intended to serve as the final principal payment, or as a payment source if government transfers are delayed. That fund was recognized in the financial statements as a separate item, not as part of the state's financial debt. The valuation as of December 31, 2025 states explicitly that the final payment is already in the company's hands, which is why the state's remaining debt to the company is described as NIS 200 million.

That chart matters because it shows why the bond can feel almost sovereign at first glance. At year-end 2025, Series A still had NIS 250 million of principal outstanding. Of that, NIS 200 million reflects future transfers the state still has to make under the agreement, while another NIS 50 million already sits inside the reserved fund. Economically, that is a very strong wrapper. But it is still a wrapper that runs through Amidar.

The balance sheet tells the same story. At year-end 2025, Amidar reported the state's financial debt as about NIS 100 million current and about NIS 95.8 million non-current, together about NIS 195.8 million. That is not the nominal amount, but the balance-sheet value of the amount still to be received. In other words, the financial statements recognize a financial asset owed by the state to Amidar, not a direct claim of the bondholders against the state.

What The Agreement Does Not Give Bondholders

This is where the line that matters is drawn.

First, note 18(c)(1) says plainly that the debt-raising agreement is between the government and Amidar only, and does not grant rights to any third party. That is unusually explicit language. If the state wanted to create a direct guarantee to bondholders, this is exactly where you would expect to see it. Instead, the note says the opposite.

Second, the same note says the agreement does not provide for state payments to accelerate if the company's payments to bondholders are accelerated. That is also material. A direct guarantee would normally be expected to move with the accelerated debt. Here the state's mechanism remains a separate contractual schedule.

Third, the reserved fund itself is not pledged to bondholders. The note states that the reserved-fund account is not pledged in favor of the bondholders. That improves near-term payment protection, but it still does not replace the legal structure with a direct bondholder claim against the state.

Fourth, the annual bondholder disclosure does not try to blur the issue. In the table for the series, the company states that there is no guarantee for the company's obligations under the trust deed, and later states that the bonds are not secured by collateral. So even at the level of dedicated disclosure to bondholders, the message is that this is Amidar debt supported by a state mechanism, not state debt issued through Amidar.

Why The Rating Still Looks Sovereign

This is where Midroog comes in, and the logic is actually very clear. The rating does not rest on Amidar's standalone strength. The follow-up report says the rating is determined under the GRI framework and without setting a BCA. In plain terms, Midroog is barely trying to judge Amidar like an ordinary company on a standalone basis. It is asking a different question: how likely is the state to support the company if needed.

The reasons are laid out directly. The state is the owner, the main customer, and the regulator. Amidar is the government's execution arm in public housing. Midroog also says that the state's refusal to support Amidar's obligations would cause meaningful reputational damage and harm market confidence. That is why, despite the absence of a direct guarantee, Midroog sets the rating of the debt at the level of sovereign risk on the local scale.

But the report then makes a second and even more important move: it stresses that the rating does not assume that the debt-raising agreement provides a state guarantee to the bondholders, and does not assume that the payment mechanism isolates the state's payments from Amidar's condition. Midroog says explicitly that those state payments remain exposed in a scenario where Amidar ceases to be a going concern.

That is the real point. If even an Aaa.il rating note needs to clarify that there is no direct guarantee and no full ring-fencing from the company's condition, then the rating is resting on a high probability of support, not on a direct legal claim of the bondholders against the state.

Midroog's wording on the stable outlook and downgrade triggers reinforces exactly that reading. The stable outlook reflects its assessment that Amidar will perform its side of the agreement and that the state will support the company to prevent a default event. The downgrade triggers, by contrast, include a lower assessment of state-support probability, weaker linkage between the company and the state, adverse changes in the management agreements, or leverage expansion beyond the agreed debt framework. Those are support and relationship triggers, not hard-guarantee triggers.

History Supports The Structure, But Does Not Change It

The most tempting place to slide into the wrong conclusion is the operating history of the mechanism itself. Midroog writes that from the September 2018 bond issuance through September 30, 2025, the state transferred about NIS 750 million to Amidar for the ongoing bond repayments, and in 2018 transferred another NIS 50 million for the reserved fund. In other words, the mechanism does not exist only on paper. It has already functioned in practice over several years.

That is strong evidence of real support in practice, but it still does not turn the legal structure into a guarantee. It only shows that the state has so far fulfilled what it undertook to Amidar, and that the rating agency treats that operating record as another pillar of confidence.

At the same time, the follow-up report adds another important detail that explains why the bonds cannot be read as though they sit outside Amidar altogether. Midroog says the state has a right of setoff against Amidar's obligations to it under the management agreement, and that Amidar's assets are subject to a general floating charge in favor of the state. The annual filing also describes a broad state setoff right under the management agreement. So this is not a bankruptcy-remote wrapper that cleanly separates the bondholders from the rest of the Amidar-state relationship. It sits deep inside that wider relationship.

What This Means In Practice For Reading The Bond

The practical conclusion is that the key question in Amidar's bonds is not whether there are enough indications of state support. There are. The real question is what kind of support this is.

If this were a direct guarantee, the main point to monitor would be the state's own obligation. Here the monitoring framework is wider: bondholders need to watch the structural linkage between the company and the state, the terms of the management agreements, the leverage framework the company stays within, and the continued state interest in preventing a failure at Amidar.

That is also why Series A should not be read as a disguised government bond, but not as an ordinary corporate bond either. It sits in between. The strong edge of the structure is that this is a government-owned company with a payment mechanism that has already worked in practice and with a rating derived from sovereign risk. The weak edge is that bondholders still rely on Amidar as the direct debt layer, without a direct state guarantee and without full isolation of the payment stream from institutional or legal risk at the company level.

Conclusion

The precise way to read Amidar is neither "there is a state guarantee" nor "there is no sovereign leg here at all." Both readings miss the point. The correct reading is that this is Amidar debt with unusually powerful state support, but without the final legal step that would make bondholders direct creditors of the state itself.

That may sound like a semantic difference, but it changes the risk checklist. As long as the linkage to the state remains strong, the management agreements stay intact, and leverage does not move outside the framework around which the story was built, the rating can continue to look almost sovereign. If one of those pillars weakens, bondholders will discover very quickly why it mattered to distinguish in advance between support and guarantee.