On-Chain Securitization Roadmap / New Silver audit


This roadmap is based on an initial review of the New Silver Tinlake legal structure by Shearman & Sterling and from our own internal monthly audit process. It provides action points that will lead to a secure on-chain securitization mechanism.


The initial DAI generated by MakerDAO for RWA was done on April, 21th and invested through New Silver DROP for real-estate bridge financing. Since then, we are now at 6M DAI invested in RWA. We enjoyed significant PR and first-mover advantage. This was instrumental to now have access to more established players willing to be onboarded on MakerDAO (Tinka for instance, with a 270M€ credit book).

While the Cayman Foundation and the 6S Trust structure are great ways to invest the MakerDAO balance sheet in RWA, those remain in the TradFi space. Ethereum is merely used as a source of capital and not used as a settlement mechanism. The work here intends to provide institutional-grade on-chain securitization. This could be a game-changer for the financial system and MakerDAO could reap the rewards of being the cornerstone of this change.


Our strategy to achieve high TVL of on-chain securitization is the following:

It will not be easy, nor fast, nor cheap. But this is the key to unlock a multi-trillion asset class on DeFi and disrupt how securitization is done in TradFi.

While the end goal is to have liquid senior tranche tokens with all the ratings and safeties you can expect from large pool securitization, the roadmap is split into three buckets. Things that we should address in the short term (before continuing increasing DC), things that are important to raise Debt Ceilings in the medium term, and finally, things that we need to disrupt achieve an institutional level in the long term.

For sure, we will discover other items along the way and this is a living document.

Short term action points

Short-term actions points are Items that are either important and need to be fixed quickly or items that are just easy to fix. This should be done before moving forward.

Revisit the Independent Director clauses

Scope: All-Centrifuge related assets
Problem: The Independent Director agreement is cancellable by either party within 30-days which makes it non-blocking and quite weak. Some Material Action clauses are also not great.
Solution: Let Shearman & Sterling (our counsel) and Manatt (Centrifuge counsel) find the good wording.

Timeline: End of 2021

Construction cash escrow

Scope: New Silver
Problem: Any fix & flip loan is divided as an acquisition part and a construction part. The construction part is liberated over time with spot inspections. Currently, this cash is escrow at New Silver Lending LLC. Some clients of New Silver are fine with that method as New Silver should be able to access funds quickly to give it to borrowers.
Solution: New Silver will move the custody of the construction loan cash to NS Pool LLC (the SPV).
Timeline: End of 2021 (expected to be effective on September, 1st)

True sale vs Security Interest

Scope: New Silver (?)
Problem: New Silver assigns the mortgage collateral to NS Pool LLC (the SPV). The SPV, therefore, is making a secured loan to the Asset Originator and not directly owning the loan to the borrower. This would make the SPV subject to an automatic stay if New Silver Lending would go bankrupt.
Solution: The preferred solution is to issue the loan directly from the SPV or to get a true sale assignment. Alternatively, it would be possible to structure it as a mortgage repurchase agreement.
Timeline: End of 2021

Security interest for TIN/DROP holders in the Tinlake Protocol Service Agreement (TPSA)

Scope: All Centrifuge-related assets
Problem: If the Tinlake License is terminated, this could lead to issues for the ongoing operations.
Solution: Let Shearman & Sterling (our counsel) and Manatt (Centrifuge counsel) find the good wording. A solution is to have a security interest for TIN/DROP holders in the Tinlake Protocol Service Agreement. Centrifuge perspective is that the software is open source already and plans to remove the need for TPSA altogether.
Timeline: End of 2021

Pro-rata exit

Scope: All Centrifuge-related assets
Problem: In DROP Subscription, section 4.E. introduces a first-in-first-out (per redemption epoch) for DROP token repayment, or at least might seem to. Such a mechanism is not present in the smart contract which uses a pro-rata basis and doesn’t take into account the epoch of the redemption order.
Solution: Change the wording of the DROP Subscription to something easier to understand and where everyone could agree that it represents what is happening on-chain.

Currently proposed wording by Centrifuge:

“(ii) any DROP Tokens for which a Redemption Request was received but not fully satisfied in such Epoch (each, a “Priority Redemption”) will be fulfilled in one or more subsequent Epochs pro rata with any and all other Priority Redemptions not yet fully satisfied.”

Timeline: End of 2021 (effective for new subscription and the MIP21 docs will/could be updated)

Limitation to change the executive summary

Scope: All Centrifuge-related assets
Problem: In DROP Subscription, section 4.F. introduces a wide possibility for the Asset Originator to change the terms.
Solution: Incorporate some fixed terms (that can’t be changed) and introduce a floor to the DROP APR linked to a risk-free rate + a spread. This will be deal-by-deal specific.
Timeline: End of 2021

Comment by Centrifuge

We are open to adding certain fixed terms to the DROP subscription agreement however we don’t believe that this presents any material risks today as DROP investors always have an option to reject any proposed changes during the review process that was implemented with the independent director.

DROP and Investment Company Act

Scope: All Centrifuge-related assets
Problem: DROP can be seen as relying on the Investment Company Act and need to comply or have an exemption
Solution: Looking to fulfill the rule 3a-7?

Medium-term action points

Those items are key to increase the Debt Ceiling for on-chain securitized RWA and get early traditional lenders using on-chain securitization.

Collateral Agent

Scope: All Centrifuge-related assets
Problem: In securitization, a collateral agent is usually appointed to hold a security interest in the assets of the SPV. This ensures the protection of the investors and increases bankruptcy remoteness.
Solution: A Collateral Agent/Trustee should be appointed to hold a security interest in the assets of the SPV. It is not clear at this stage how that would work with the NFT-locking under Tinlake. Maybe the Collateral Agent should be the operator of Tinlake? It might also be solved with the current plan from Centrifuge to orphan the SPVs and make them run by Trustees.

Improve Tinlake design for non-fixed-rate or with cashflows credit pools

Scope: All non-fixed-rate credit Tinlake pools or assets with cashflows
Problem: Tinlake was designed with invoice factoring in mind where there is a fixed rate and a bullet payment at the end. An example is the issue of managing payments for New Silver, the fixed interest rate for revenue-based finance (15% for Corl where it is expressed in % of revenues), the convoluted way to invest in farmlands.
Solution: ?

Improve MIP21/MIP22 design

Scope: All RWA assets
Problem: In the current design MIP21, cull is working in a binary fashion. Either vault is fully collateralized, or the collateral is worth 0 DAI (and the bad debt is subtracted from the Surplus Buffer). This makes it risky to have a MIP21 Collateral Debt Ceiling above the Surplus Buffer.
Solution: Make cull more granular.

Back-office/Accounting for Tinlake

Scope: All Centrifuge-related assets
Problem: As DeFi is not yet understood by accountants, it is difficult to audit the SPV accounts.
Solution: Accounting documentation under GAAP would be a start. Probably having a way to extract ledgers from Tinlake would help as well.

Comment by Centrifuge:
While most of this information is available on-chain it is tedious to get it from there. We are continuously improving our subgraph (GitHub - centrifuge/tinlake-subgraph) which will include a lot of the necessary details of the on-chain transactions. We are looking into how to reconcile this with the bank account side in the near future.

Transactional safety

Scope: All Centrifuge-related assets
Problem: Tinlake requires the use of wallets and DAI transfers, a lot of mistakes can be done (divulgation of the private key, mistakes in the DAI transfer, …). While this is a usual problem for DeFi, we should fix this to reach DeFi for RWA scale.

Comment by Centrifuge:
We are addressing this in a future release with a multisig that limits which addresses can withdraw DAI to a set of pre-approved addresses. Using this dedicated smart contract wallet together with a licensed BD for the fiat on/off-ramp. This greatly reduces the risk of loss of funds. A spec is available publicly here: Multisig Tinlake Proxy - HackMD

Independent SPV

Scope: All-Centrifuge related assets
Problem: Currently, Centrifuge SPVs are member-managed by the Asset Originator. While this provides significant flexibility at a low cost, this construct will not scale with higher Debt Ceilings.
Solution: Centrifuge has a roadmap to create a trust to manage the SPV where the Asset Originators will have only limited powers.

Provide legal ownership for MakerDAO DROP tokens

Scope: All-Centrifuge related assets
Problem: Currently, MakerDAO is lending against over-collateralized DROP tokens in custody in a smart contract.
Solution: A MIP58 Cayman Foundation (or any orphaned SPV controlled by MakerDAO) can probably solve this issue. The hardest part will be to change the smart contract.

Long-term action points

This list is about making on-chain securitization equal or better than off-chain securitization. This will allow all institutional lenders to invest in on-chain securitization.

Credit Rating Agency

Scope: All Centrifuge-related assets
Problem: Scale at managing DROP investment would be quite easier if we could make them more fungible by having a credit quality rating by a recognized credit rating agency.
Solution: Develop a rating methodology with rating agencies. Implement preconditions from such rating agencies.

Lender voting mechanism/administrative agent

Scope: All Centrifuge-related assets
Problem: Currently, the DROP tokens are passive and can’t communicate on-chain with the Asset originator. Even with the addition of a collateral agent, there would lack a way for the lenders to direct it. Usually, lenders should provide their wishes in some ways (by vote, unanimity, or through a Lender Representative). Having an on-chain method would greatly improve the off-chain method currently used in TradFi.
Solution: Develop a voting mechanism for DROP/TIN tokens. Elect an administrative agent (Lender Representative) that could take care of DROP administration for the benefit of the lenders (and why not having MakerDAO directly or indirectly assuming this role as the cornerstone of DeFi).

Yank the bank mechanism

Scope: All Centrifuge-related assets
Problem: Currently, the executive summary can be changed and a negative action should be done by investors. Usually, in TradFi, with the absence of a positive affirmation, the solution is to yank the bank, i.e. buy back the DROP tokens of the investors not willing to pursue with the new executive summary.
Solution: For significant changes to be implemented, an affirmative vote will be needed and all investors not issuing the affirmative vote should be repaid and excluded from the pool before the change.


A number of these look fairly important to fix before the end of 2021 — especially the ones that have been a familiar point of contention for months now.

Given that we are footing the bills to improve a legal structure that should have already been 1) reviewed by us prior to involvement, and 2) developed more rigorously by Centrifuge, I would like there to be more public discussions between Maker and Centrifuge on prioritizing and fixing each of these points.

I am very glad to see we have finally gotten a legal review of this structure after already approving $44 million in exposure.

The extent of the problems make me leery of on boarding more Centrifuge assets without a firm commitment to begin fixing some of these immediately.


Did we ask Shearman & Sterling to write an enforceability opinion on the Centrifuge docs?


So I asked Shearman to audit legal documents and tell us what to do for the benefit of MakerDAO. I’m no legal expert, they are. So I report what they told me after reviewing this exact document with them. If they were in the view that we should get an enforceability opinion (and that should come from the borrower in my view) I would have written it. But I can ask them about it next time.

To me what is important in any deal with the legal document is that both parties (buy-side and sell-side) have a lawyer that is working for their benefit. I know that you are in the view that having some enforceability opinions removes the need for a legal review but that’s a point where we disagree. A code with 80% audited can still be buggy in the last 20% and the code might not make what is in our best interest overall to start with.


Sounds good, thank you.

I agree that if someone doesn’t feel comfortable reviewing a document, they should hire a lawyer. For me personally, since these documents are not terribly complex (IMO), I’m comfortable just reading them and feel like I understand our rights. But any analysis whether it’s done by a lawyer or an amateur like myself, needs to be predicated on the knowledge that the agreements are valid and enforceable. So that’s why I think these documents are so important.


Then why was this not done months ago?

1 Like

As answered during the call, because we didn’t have lawyers. I think I was quite transparent that first it should have been done by a Legal CU that didn’t see the light, then as we think it is an important issue we asked a budget to get lawyers and then having to find an interested one and pass KYC. I think my reporting provides enough visibility in what I do.

And you made those clear on the forum (“piss poor structure” or something similar). You insisted to have an independent director quickly and it was done not right. Sadly we didn’t hear you saying it wasn’t done right and providing reasons (and I shared the document before implementation as you wanted to review it).

I fully hear the vocal people here, but my job is to act for the benefit of the DAO and to do what MKR holders want. And they passed the Centrifuge collaterals so far. My job is to facilitate not to make decisions. We highlighted some issues, some community members highlighted others and it was clear to everyone that there was no legal review. Under which mandate should I go against MKR wish?

If some MKR holders are not happy with what is presented to them in on-chain votes, let’s chat and we can adjust what is presented to Governance.

1 Like

No one is asking you to violate an executive spell.

The problem is that your core unit also does the risk assessments for these. It is the job of your CU to point out and emphasize dangers. If you don’t want to assess risks and present an opinion on how to protect the protocol from losses, please have @Risk-Core-Unit do the risk assessments.

1 Like

@PaperImperium , isn’t this post presenting risks as they are today, and how we are going to work to solve them to protect Maker?

All core units are here only to facilitate MKR decision-making.

I don’t understand your point.

No. It’s presenting risks as they were back in April.

The only risk that is less today about this structure than in April is the existence of an independent director at the SPV level — and only after the repeated insistence of @g_dip and others. And apparently even that was done in a way that was “quite weak.”

From the RWF CU mandate:
“Use real-world finance to produce more yield while keeping our risk appetite low thanks to wider diversification and risk management tools.”

Keeping risk appetite low thanks to risk management tools. I think your CU is supposed to have opinions on these RWA that come in front of us :man_shrugging:

But what is important is that Maker now has an actionable to-do list! So let’s start checking these off so that we can continue to onboard collateral from Centrifuge without worrying we will end up as a headline on Rekt.news

I would like to see an independent third party within the DAO who can observe this process.

1 Like

Thanks for sharing this with the community, it is extremely helpful. It covers many structuring points, but not risk alignment points with the other stakeholders. I have been trying to find more about risk sharing with the originators and extra security liens like hard assets collaterals, programmatic buyback of non performing positions, or pre agreed NPL sales to credit collectors. However I have not found much. Is it due to my inadequate skills of navigating the forum or to the fact that those are confidential info not shared with the community?


Welcome aboard the Maker Community Luca! We have not crossed the nonperforming loans chasm–so far no unpaid principal balance/ delinquencies. Very early in the game, but totally-you have brought up some good questions. TY, and again, benvenuto

Thanks for the welcome. In New Silver case, where credit is not too granular, structurally pro cyclical, and back-ended in nature, it is even more important to have all the incentives aligned. We want to be considered ‘the smartest money’ by investors. Looking forward to hear more on those topics. :pray:t3:


@SebVentures nice to see the report with actionable items.

I think we really need to limit the total DC for NewSilver/TinLake until a number of these issues are taken care of.

I think this is important the ‘Maker is footing the bill’ part here to solve the issues we think we are stuck with. Literally governance can DC/0 existing facilities to not allow them to extend until a certain critical path list of items are dealt with. I am NOT suggesting this needs to be done on my first round reading. But if anyone came up with a good reason for doing such I would not stand against it.

I feel like we just plowed ahead with NewSilver rather than wait because everyone wanted RWA to come on line to solve the DAI supply problem. Yet we have another key player in the space that as far as I can see has dotted almost every i and t - still waiting for a DC - 6s. Unclear to me what to make of this, I have my own opinion and will leave others to form their own. I do think process should come to question though for many reasons brought up in the thread I will not repeat here to prove the point.

What I am looking for from RWA CU here is a risk assessment regarding ‘critical path’ items in the list. I don’t see a ranking of these issues on ‘any scale’ that gives me a feeling for criticality to the system or our funds and whether any form of actions should be taken.

I also really would like an opinion from RWA CU unit regarding limiting DC - a specific recommendation in that regard.

I really want to echo this and wonder does each CU have a document delineating their authority and responsibilities? If Maker CU’s don’t have these I would urge them to put together some and present them for approval. Given the intimate nature of RWA requiring NDAs etc. and needing specialists to even analyze this stuff I would have thought only the RWA CU team and their legal consultants ‘could’ do these risk analyses. I don’t see how the risk CU could - or they would have to hire a separate unit specialized on RWA to even have hope to doing this.

Given Maker is already partially in here I think cutting up the actionable items into critical path - limited DC until done, whether and what that limited DC is, vs. not critical path would be a prudent next step so the tasks can be lined up. I also hope that if Maker is paying for this work that the costs associated are broken out. My biggest concern here is the negative PR of an article suggesting Maker is paying for 3rd party Tin Lake legal expenses to straighten out their RWA token legal structure to the benefit of New Silver and detriment of MKR holders.

The short terms items should be fixed quite quickly and I don’t think we should onboard/increase anything before most are fixed (following whatever expert will tell us). I don’t think we should reduce DCs as everyone is working in good faith to fix those issues asap.

Having a Collateral Agent/trustee is the most important thing to me outside of the short-term section. It might or might not be easy to have. The main issues are who directs the trustee and how does it work with the NFT locking.

Ownership of the DROP is also one item (not in this list as not related to Centrifuge). The DROP is owned by an “autonomous” smart contract (MIP22). The Cayman RWA Foundation is probably providing a solution to that but will require some work as well.

Our investments in RWA are paying as we see more sophisticated asset originators coming to us and people with a financial background interested to work with us. This will make RWA on MakerDAO (and more broadly in DeFi) safer. There is a kind of feedback loop.

The pace is defined by MakerDAO. This is why I asked earlier this year and the answer was quite one-sided toward growth and leading RWA in DeFi. We can also be more cautious and onboard only things that have a full pristine legal review by MakerDAO lawyers (and get those paid by the borrower) and get an anchor institutional co-investor. That would delay all RWA until we get a Legal CU and would require RWA to already be mature in DeFi (which was the slow path in the post).


My general opinion on this that I stated periodically.

1% of outstanding or no more than 50% of surplus buffer DC on any one experiment is probably tolerable. This currently works out to this 25-60M range. It is a loss that would hurt but would not break Maker. It would teach us a lesson as well. I keep thinking 10M isn’t a big deal and we can take on 3-5 of these without breaking the system as it spreads risk. Provided SF was also reasonable (2-3% range) it would provide income (.75-1.8M/yr) that could have been earmarked specifically to hire outside legal opinions.

There is basically a set of optimization loops going on here that we need to keep in mind. I literally want versions on these legal document structures as well as authors because given the newness here I expect these structures will undergo multiple revisions for reasons that will come from various parties over time and I really want to know document structure inheritance as well.

It is my impression from the Maker side that we really want and need RWA to ramp up, and to kick this off Maker governance (with polls, executives etc.) is more prepared to take on the credit side risk against the loss of DAI vs. doing extreme DD and holding this back. Our only issue is whether we feel comfortable enough to extend DCs. I think given your above report on Centrifuge there are some issues that need to be dealt with before we extend their DC(s).

Again what is critical path here to a DC raise? Personally I am concerned we need to tap on the brakes with Centrifuge/New Silver and all the new DROP vaults they want, and let 6s get up and running. In the loosest sense diversifying risk in this RWA arena means looking to bring on more RWA players looking for 10M so we can diversify not portfolio risk but RWA counter party risk as we ramp up legal opinion, risk assessments, and onboard qualified people. This will cost more initially and probably place a high cost demand on legal opinions, risk assessments, manpower, etc. but if Maker is going to wave past being more careful on the first 10M RWA borrowers, well lets spread these intial 10M risks out to more RWA players and focus on what we will need to do to get the first RWA player from 10-20M DC to the next step 100M DC.

If New Silver/Centrifuge wants this - well then they need to start fixing the issues we are seeing. If this takes a while - so be it. I want to be more protective of getting to 100M because next step from there is 500-1B and by the time we hit that number we better have our collateral custody chain (liens) well in hand.

The race here isn’t who will be first, but who will be best. Whoever wins this is going to be first to 100M facility and they will have laid the hard ground work easy to follow the rest to follow.

The RWA job here is not just to facilitate getting somewhere, but to make darn sure risk is covering this, and legal has some handle in here as well and to start lining up funding, legal consultants, and people (and god knows whatever else we need - Eng Dev, smart contracts, PR, biz, ??) to get this ramping hard and fast.

While Maker has other issues on the plate I see RWA as priority one effort until we see that first 100M or better 1B vault. Remember our RWA players are going to need time to ramp business as well.

My question to New Silver/Centrifuge or any RWA player is whether a pause in DC raises is reasonable and expected result of this process, or whether it will put undue strain on their plans, because if they need to move fast then Maker needs to be prepared to move with them if we can. As a delegate and MKR holder I also want to make sure we don’t get ahead of ourselves. We should not allow our desires to rapidly grow DAI and demands from RWA players take a front seat to the health and security of Maker and of DAI. I personally put health, security and sustainability as priorities over growth.


I have to echo this sentiment. Maker does not have an equity stake in Centrifuge. Centrifuge does, however, enjoy backing from quite a large number of VCs, institutions, and prominent individuals.

If JST Capital, Mariano Conti, Stani Kulechov, Julien Bouteloup, Fenbushi, Moonwhale, Fintech Collective, and other sophisticated, deep-pocketed Centrifuge investors aren’t willing to fix the legal structuring of Tinlake pools, Maker should not be doing it.

I will oppose any new Centrifuge assets until there is progress on the most urgent items on this list. If that’s the end of 2021, then that’s the end of 2021. We’ve rushed into this situation for $40+ million in exposure. Let’s learn from that, and not rush the next tranche.

1 Like

:sweat_smile: I think Mariano is calling you out Sir


For sure–one for All, All for one. Together we stand, divided we fall, right? If not, then what…

1 Like

Maker is not the piggy bank for Centrifuge or its investors.

What next, pay the hundreds of thousands that 6S spent on their structure?

It’s one thing to work around the edges for bespoke changes. But this is beyond ridiculous.

That we got to this point at all means that our RWF CU has failed Maker. This is unprotected financial sex, and makes us look like amateurs who can’t even be bothered to drop a few thousand dollars on a lawyer when tens of millions are at stake.

Someone savvy will actually read these documents and see the asymmetric power relationship. And then what? We ask nicely they not rug us?


You have a right to your opinion–I personally would not bring-up individuals personal investments to make a point. But that’s just my 2 Conti’s–you can do as you wish. ¯_(ツ)_/¯

1 Like