SES Project Real-World Sandbox


I did want to mention/discuss one key topic: how will Maker determine its credit standards and the other key terms, such as advance rate, loan tenor, pricing, for lending against various types of collateral?

Will Maker require that most/all collateral loans be structured investment grade (BBB or better)? How will credit standards be enforced?

For example here is the Eurosystem haircut schedule. They require that any collateral be at least BBB.

@Eumenes thanks a lot for your message,

We will present the high level findings of the report during this week’s SES update on Friday. The presentation will be on a high level for everybody to benefit.

Following that presentation, we will release the report in the beginning of the following week. I believe many of your points will be addressed, but I’m looking forward to receive feedback from you following the release.



I look forward to seeing it.

I think a one year probability of default corresponds to about a BB+ credit risk. (I helped build thr DBRS idealized rating transition matrix long ago)

One other point - the probability of a bond taking a 1st dollar of loss (which I mention above) doesn’t capture all the risk as it ignores loss given default.

One final point. While we need to quantify risk, its very difficult to measure 1% default risk accurately. We don’t have enough observations, the observations arent homogeneous enough, default rates are time varying…


Fully agree.

EL should be a core metric, but given that we really do not want to incur into a liquidation/ repossession situation, PD should be in my opinion our main metric to focus on.

If we do not have observations, as it always happen, we should try to underwrite risk using closest comps. That is often available, and basis of good risk underwriting. And that is where structuring becomes the most important aspect.


Makes very good sense.

One could also apply loss given default info to risk weights. If we use risk weighted capital surplus

A few observations: under the Arranger Model there will need to be strong credit and structuring expertise at the Arranger to create deals/collateral that meets the standard which is basically investment grade.

Maker will also need similar skills to specify and review that Arranger deals meet this standard.

Also Maker positions will almost all need credit subordination/equity below them to assure they meet this credit standard. Will this mainly be from a TIN like tranche or also elsewhere?

Dear all,

On the 24th, we will present the key findings of our Project Real-World Sandbox in a call hosted by SES @ 4pm UTC

The call will kick-off a consultation phase over the report, that will be published immediately before or after the call. The world doesn’t wait for consultations I know but given the importance of the topics at stake, and the power we might have as a community if we agree on the core principles of doing business with real-world counterparts, we think this period will be very important.

We hope many will join and ask as many uncomfortable questions as possible.

Thank you and talk soon, Luca


Project Real-World Sandbox: Full Report

Thanks to all the people who participated during today’s call, hosted by SES. Here is a link to the presentation. It has been a pleasure discussing the key findings of the report and (try to) answer your questions.

As discussed during the call, you can find the full report through this link. You will notice that the cover says draft for comments, and it is because the publication opens a consultation phase during which we hope to collect as many inputs as possible from the community.

Thank you again for the support so far, although we are only at the beginning.

Regards, Luca


Given the importance of the topic, I’d like to encourage the team @Real-World-Finance to candidly contribute to the discussion. Be respectful, open minded and constructive :slight_smile:. This is a collaborative hustle effort.


Thanks @luca_pro for the ‘RWA Manifesto’ that sets out an unifying vision and practical strategies that MakerDao could adopt to scale RWA collaterals safely and quickly. What I particularly like about the manifesto is that it is more like a common law rather than a civil code. Having lived under both systems (U.K. and Vietnam) I believe that ‘common law’, albeit not always seems efficient, would be most appropriate given the early stage of the RWA strategy. The principles of decentralisation and equanimity of counterparties before MakerDao reinforce our confidence and hope that by partnering with Maker we could change the world together.

At this point, I would like to have a few observations:

  • the ‘arranger model’: hopefully we can see it being fleshed about a bit more, perhaps not necessarily in the document, but as more and more arranger ‘case law’ appear. For example, when referring to ‘skin in the game’ the report said the arranger might not necessarily need to provide an extreme high level of risk capital’ - What is ‘extremely high’? as a proportion of total credit enhancement in a particular structure?

  • how to reconcile the focus on short term corporate debts (eg invoice financing) and ‘bank’ loans, among others, and not on ‘long term’ corporate debts with the ‘a case for Clean Money’ where collaterals such as corporate green bonds (use-of-proceed green bonds are often long term corporate debts) are onboarded? Green financing is at a stage where longer term financing eg construction loans, are needed to bring the green assets into being.

  • Issuer countries listed have strong rules of law but how do we deal with a cross border situation, which is eligible, where underlying collaterals were originated in a third country but, for some reasons such as currency control, tax, or local banking regulation, are unable to be sold to the issuer SPV located in the likes of US, Canada, U.K, EEA, Japan? It is quite typical in trade finance, for example, that the exporter invoices are originated in a country, say, in Asia to a buyer, say, in Europe and it is not always possible or efficient to transfer the assets to the Issuer SPV in an ‘accepted’ jurisdiction such as Luxembourg.

  • Blockchain is mentioned only once in the context of RWA collaterals not being native to it. Whilst I agree that we can learn/replicate TraFi, particularly on proven legal structures that offer strong protection, isn’t the opportunity here is also to innovate beyond Trafi with the power of blockchain? starting with on-chain tracking of collaterals and covenants?

Perhaps some of the above were discussed on the call this afternoon which I couldn’t attend. Hope this provide some points for discussion as we are going through the very onboarding process at the moment. (Made some edits on second reading of the report)


Thank you for your message @quntangled. Too bad you couldn’t join the presentation, SES will post a recording of the session soon I believe so you will have the chance to go through what has been discussed.

Thank you also for the support to the core principles of our initiative, I believe this is the right spirit to enact change. Let me try to respond to your points.

@rune post on the arranger model was a powerful call to action to the community, that in some ways brings back concepts that have existed in the world of real-world credit already. What we mean when we referred to a model open to arrangers in the report, was that our credit framework is agnostic to the existence of arrangers. In our mind, arrangers are facilitators, intermediaries, that leverage their knowledge of Maker to facilitate business willing to access. They are, in their purest form, for a lack of a better term, consultants to the counterparty. In their role as consultant, they might not be required to put sizeable capital at risk. This doesn’t mean that capital at risk shouldn’t be put in, it simply means that such capital might come from other counterparts, i.e. the ones that are truly involved in stacking the loan structure. In the case of arrangers that are not purely consultants but a hybrid form, the situation would be different. Again, our framework wants to be agnostic vis-a-vis structures and counterparts, at the end of the day that junior capital at risk must be there, and those who do origination and retain the management of such risk should remain heavily exposed not only to the upside, but also to the downside.

The appropriate skin in the game will depend on the structure and on the underlying collateral. As a rule of thumb, in the market out there it might vary between 5% (minimum securitisation residual requirement) and 20%+ (for novel or riskier exposures). Maker exposure should be senior to that and, when the conditions require it, senior to a mezzanine tranche that would be provided by creditors with higher return requirements.

This is a great question. In the report, we mentioned the credit types that are the most appropriate for Maker’s set up. That doesn’t mean that those types should be exclusive. What this means, is that other types of credit would have certain structuring and risk mitigation requirements that allow them to be digestible by Maker. Also, what it means is that those characteristics should be taken into consideration when a portfolio approach emerges, i.e. when Maker would put together its risk appetite for what % goes into revolving credit, what % goes into long-term green bonds, etc.

Another thing to note, and something that I should actually clarify in the next version, is that the report refers only to credit risk and not to anything else. A portfolio approach should be considered also to handle liquidity and ALM (asset-liability management) risk.

This is a legal point that I am not the best person to answer to. But a crucial point to assess.

Great point, and something we have discussed in the call (towards the end). I personally believe that in the long term every aspect of the credit flow will migrate into the blockchain, and can’t wait to see this happen. But it will be us who will need to drive the migration, and a strong mental model is crucial in setting the right direction.

I hope I answered your questions, looking forward to more comments.


MakerDAO as a central bank

I understand that the author wants to place MakerDAO in line with central bank in term of behavior. One thing is to have a meme of decentral bank of DeFi the other is to have a serious discussion on this. Central banks enjoy a monopoly and are not driven by profit. Even when the bank of England was for-profit, shareholders weren’t well treated. Is that what we want for MakerDAO?

Indeed, it cannot be wondered at that the Bank proprietors do not quite like their position. Theirs is the oldest bank in the City, but their profits do not increase, while those of other banks most rapidly increase. In 1844, the dividend on the stock of the Bank of England was 7 per cent, and the price of the stock itself 212; the dividend now is 9 per cent, and the price of the stock 232. But in the same time the shares of the London and Westminster Bank, in spite of an addition of 100 per cent to the capital, have risen from 27 to 66, and the dividend from 6 per cent to 20 per cent. That the Bank proprietors should not like to see other companies getting richer than their company is only natural.

Bagehot, Walter. Lombard Street : a description of the money market

It is also unclear how much of a monopoly DAI enjoys. Sure, it is currently the main decentralized stablecoin but it would be hubris to think it will last forever. This hypothesis is not backed by research in the report but is used as an important assumption.

Maybe the report should explore more in detail why MakerDAO should be seen as a central bank and the implication that.

Risk appetite of MakerDAO

Unrelated to consider MakerDAO like a central bank is the risk appetite. And it could be in line with central bank (which are banks after all). @luca_pro suggests mimicking the ECB risk appetite.

First, the logical conclusion is to keep RWA small for the time being. If the aim is to generate a few basis points, maybe doing nothing is even better. Luca asks for less than 1% of PD (probability of default). Assuming loss given default (LGD) is not 100% and market are somewhat efficient, it will give you a few basis points over the risk free-rate.

Notice as well that the ECB will accept some collaterals but will add a haircut (a CR). You can see that for Asset-Backed Securities (what we are dealing with mainly), you need to be at A- or above and still end up with a 24% haircut (CR 130%). This 24% of capital should come from somewhere.

The short term ETF proposal was already riskier.

There is no discussion about the cost structure. Going super safe and getting 0.05-0.20% interest rate could lead to a certain loss after factoring in the structural costs.

In our current pipeline, no one besides SocGen is even rated (Tinka might get a rating on a similar ABS issuance tho). So if the community wants to go that road, we can start by putting the pen down and clearing the whole RWA pipeline. How do we create a new pipeline for such a risk appetite?

It is my view that this proposal of risk appetite is not great from a business perspective (not earning anything). Moreover, the report doesn’t provide any indication on how to source such collaterals and who the borrowers might be.

I worry that the Arrangers will be seen as “banks” and that the haircut will be lost in the discussion (more precisely to match the reality). We will go from super-senior central-bank-like type of risk to junk dressed as a super senior. Junk is super easy to catch, it’s anything that gives you more than 2% currently (it is not bad by itself, Coinbase is rated junk). Super senior (in the understanding of being central bank-like) is below 0.25%.

On the structure

I think the hierarchy between RWF Compliance and the RWF Portfolio CU is unusual. Shouldn’t it be on the side? The fact that RWF Portfolio CUs are incubated in the Compliance CU is discouraging for diversity. Diversity of RWF CUs is needed to provide a diversity of opinions to the MKR holders to make an informed decision. It is not decentralized if one CU is a gate.

I wonder if we shouldn’t define RWF Portfolio CUs as credit rating agencies and require many sets of eyes for big DC. You can easily be compliant to a framework but underwriting garbage. To the extent that the credit policy is to lend only against high-quality collaterals, I’m not sure we need such CUs at all (or not too much). We already have credit rating agencies.

My belief is that MakerDAO should be lean and limit governance. Not to create a wide bureaucracy. It is not possible for RWA currently, but shouldn’t we move toward a governance minimized RWA structure instead of a crawling bureaucracy? Your stated risk appetite (central bank-like) is quite supportive of a low headcount anyway.

If we want to underwrite exotic issuances (i.e. anything not rated investment-grade and coming from a flagship financial institution), we should have a discussion on why we would want to do it? Is it because we have a better acquisition channel? Better cost structure? Smarter than the rest? We did it so far as there was no alternative, due to the MIP6 mental model and because we had to start somewhere. We should probably challenge those assumptions.

Regarding the RWA Legal CU, I’m fully aligned. “How” is the only question.

Regarding your last comment on the Arrangers

From @rune post:

Specifically Arrangers have the autonomy to onboard deals based on Boxes that represent a pre-approval for a particular Elibility Criteria, and that each have unique terms.

The Arranger is the seller to Maker (with autonomy to sell and an economic incentive to sell) and it is sound practice to ask for 5% risk retention from the seller. I understand it is not possible with most Arrangers of the lineup. Should we breach this rule to match with the reality? Maybe. But we should be clear on what we are doing.

We can’t say we are ultra-safe like central banks then using arrangers (all startups) not putting 5% skin in the game and originating non-rated assets. I think the report should tell MakerDAO community how to reconcile those two extremes (or help choose between).


Thank you for your detailed email @SebVentures,

The importance of your experience in structured credit for Maker in the last months cannot be understated. I will try to answer your questions based on my perspective after 15 years in the business but I am happy for other members of the community to chip in. I also believe those points valuable and I’d love to have an open debate at some point in the future in a call format as well.

Thank you for reaffirming the spirit of this conversation. Far from being literal, the report intended to show a direction of travel that, as we said, resemble more that of a central bank rather than an investment bank. There are several similarities that Maker enjoys with central banking, but I really don’t want to spam the community with the rants of a monetary economist more than what I do already! The report was not intended to be an educational piece on monetary policy and central banking but I can affirm clearly that I strongly believe that an expansionary footprint of the DAI is conducive of the sustainability of the DAI as relevant stablecoin in the ecosystem. And that an institutionalisation of the real-world asset framework is a great facilitator of such expansionary footprint. This is, obviously, my personal view based on my experience, and I accept and respect other people’s opinions.

There is one important thing, however, that I would like to state. With no delegation of power away from MKR token holders, I believe that those same token holders are the best placed to protect their own interests when approving or rejecting an application, also based on the qualified opinions provided by the CUs. For what I can say, I confirm my view that this direction of travel is compatible with the maximisation of value and the fulfilment of the mission of Maker. I have been happy to see that many, within the community, who own significant interest as token holders, have shared the same view.

On the contrary, I strongly believe that an institutionalisation of the approach, and a revision of the risk appetite, will massively expand the initiative from what it is today. As you can see from the first page of the report, such report was endorsed by few large originators that already interact with maker (@mrabino1 from 6s, @Jason and the rest of the END-labs team, @martin and @roollie and the rest of the Centrifuge team). I have also had the pleasure to receive several calls expressing interests in participating to the initiative from former business connections. Among those interested parties, that I cannot share at this stage but that I would direct to the appropriate Maker onboarding units, there are 2 sovereign funds, 3 investment banks, 3 alternative originators operating in the invoice financing space, 2 large credit funds. All ready to provide credit quality that is way within those parameters. Such institutions would not be ready to approach without a structured organisation that “speaks the same jargon”. That is why I am so enthusiastic to unite forces and collaborate to speed up the initiative.

For what concerns the return on capital, I believe there is a bit of confusion here. The private credit market is not the public credit market. There are several very high level credits out there that are not enjoying the size and the access to public credit through bonds or other structures. Although the implicit cost of credit at the originator level is very small, those counterparties still pay few % points above swap rate because they need to fund the whole structure. To put it in simple terms, although the cost of liquidity for a bank is 0% nowadays, its cost of equity is 10-12%, and that means that such cost of equity needs to be compensated by the spread charged to customers ultimately over the cost of financing. It is such cost of financing that constitute the hurdle for Maker. When I was board observer of an ECB bank, we could solidly generate 3-4% from IG private credit originated through platforms. I believe that within Maker, with lighter regulation but post Covid, we could do the same at 0.5-1%. I understand that many of those concepts could be intimidating, and that is why I believe part of the job of the CUs should be educating and discussing all those concepts within the Community. I know that @williamr has a keen interest in giving back through education, and I am so happy to join him even if it’s hard work.

I believe there is a bit of confusion but in good faith about this. Although the Portfolio CU would be functionally dependent from the Compliance CU, their hierarchical dependency is both towards MKR token holders. This is the same that happens in traditional banking for Audit Units. Although an Audit unit is functionally under the CEO (typically) or COO, its hierarchical dependency is towards the Supervisory Board. The employer is the CEO, but the people Audit reports to its the Supervisory Board. That is why Audit always has a seat on the Supervisory Board. The same happens with Compliance, and the same would happen for the Compliance CU. All units will be dependent from governance and effectively parallelised. I apologise if this wasn’t clear in the report, and thanks a lot for flagging it. I will amend the next version.

This is the point where I am afraid I have to disagree with you. I strongly believe that the current set up of the RWF CU is non-institutional, as it concentrates too many responsibilities and lacks transparency. I also believe that, for this reasons, high quality institutional lenders would not approach. I was one of them, and I wouldn’t. Adverse selection is real.

I am sorry but I don’t understand your comment. I believe that the report is quite clear on this point and fairly aligned with what you, rightly, suggest. I am sure I could have written it better if there is still confusion. We had this conversation during the call and I believe people were satisfied with the explanation, I should have clarified the point further. My apologies. For the interest of the community, I report one passage that is very important from the report.

Arrangers through the lenses of credit risk onboarding. This means that arrangers should be, in all shape or form, transparent in credit risk terms. Although they act as an extra layer of safety and efficiency, any collateral onboarding application they bring forward will have to go through the same level of scrutiny. Arrangers are a very powerful resource within Maker’s set-up, and should be compensated accordingly, but without compromising the level of diligence of the RWF Portfolio CUs.

As @rune mentioned, his initial post on The Arranger Model was an initial post that would have stimulated the conversation and floated some high-level ideas he thought about in pointing to a new direction for real-world finance. The Sandbox report intends to incorporate and augment this vision, and naturally would have few inconsistencies with what was said before. As you can see from the cover of the report (font was small I know!!!) @rune has endorsed the report.

Thank you so much for having spent the time to provide such detailed comments as Facilitator of the RWF CU. I hope that my answers have been satisfactory, but I am looking forward to organise a call as soon as everybody has digested it. I am sure such report is just a first step and these conversations can improve what we are doing massively. Opening the table for a structured conversation has been the reason we started the Sandbox in the first place.


This is a great report that brings together a lot disparate discussion and threads, thanks @luca_pro for putting together in one place.

The discussion above about re risk profile vs expected rates of return hits upon something that I don’t feel came across in the report yet: what is the value proposition(s) of Maker to the borrower market beyond cheap capital?

For the current borrowers and early adopters (mostly start-ups) it was probably something along the lines of “cheap credit line probably with fewer strings attached than I’d get in the tradfi market and it’s kinda cool to do something with DeFi and as we’re a start-up, why not”. This report makes clear that this is no longer the direction of travel (rightly imo).

What borrowers are NOT getting from Maker is:

  • Easier access to credit (Maker integration still a lot more complicated than getting financing from tradfi institution, even with an intermediary Arranger model)
  • On-chain financing (while the capital comes from DeFi, it’s taken off-chain via the Arranger, so actually the borrower doesn’t benefit from any form of on-chain efficiencies and transparency in their capital flows)
  • Capital with higher risk-appetite (ie funding for projects that are hard to fund in the tradfi market)

So if the Borrower doesn’t receive any of the above, then there needs to be something compelling they do receive. Atm this is low cost of capital, but is it the intention that this will remain Maker’s only advantage for the foreseeable future? If so, then any discussion around the risk / reward weighting should be framed on this basis. Maker will need to keep its cost of capital super low in order to compete with other institutions to get capital out of the door in the current competitive climate, otherwise we won’t attract the quality we want. That means accepting very low returns, which I feel has been implicitly accepted but maybe not explicitly stated yet?

What I’d like to see is an analysis of target returns and exactly how / why these will be achieved, bench-marked against what comparable tradfi institutions receive. Putting some model numbers down as a straw-man might might help the community engage with this topic and work out exactly what returns are / aren’t acceptable.


Thank you for this @justin1066, in my mind those are all great questions and points that are still open and will need to be analysed following real-life discussions with the originators on what is or is not doable.

In my mind, these would be the advantages for a borrower of any kind when interacting with MakerDAO:

  • Attractive cost of capital vis-a-vis the alternatives; often the alternative is not a bank, because a bank has other issues sometimes that make this unviable; the alternative is alternative credit providers that receive their liquidity from credit hedge funds, family offices, and venture capital funding - all with much higher IRRs. Maker could be an incredibly powerful alternative source of capital
  • Higher quantum; Maker (liquidity and risk) risk appetite will be the bar here, but I believe that the quantum that Maker can deploy is meaningful in the market
  • Potential to on-ramp into DeFi; I think your point of on-chain financing is super meaningful. It is my view as I stated many times also during the presentation that Maker can be the most powerful institutional ramp into DeFi, and that borrowers won’t need necessarily to bring back liquidity off-chain. This is particularly relevant for banking borrowers that have fungible liquidity coming from several sources. Although we should structure ourselves to allow this to happen, ultimately will be in the hands of the borrowers to choose
  • Easier access for everyone; although this is still early experimentation days, I think our ambition should be really to provide a smooth avenue to receive funding for international businesses. This might not be evident now, that we are designing and debating, but I hope this will be evident later on. Also, although I believe our core focus should be on developed market (for testing) this might not be the case in the future. In less banked jurisdictions such a non-intermediated channel could be incredibly beneficial.

I think, if potential borrowers agree, that a panel that includes borrowers/ counterparts and Maker CU members could be started to debate those ideas and understand whether there is a point of contact and how to unlock it. The MIP6 processes will go alongside it, but I think we should keep the conversation open as much as possible. Colleagues from 6s, END-labs, Centrifuge, have already provided interest. We could make it a period (biweekly) call open to everyone to chip in and help shaping the future together.

A model is a great idea. I will be happy to put numbers in an excel based on my TradFi experience, and start iterating them for the DeFi world.

In any case, I think we should all have another call in 2 weeks time, having given time to everyone to digest the report and asked question. It might be a long call (2 or 3 hours) but a precious one for practitioners. What’s your view?

Thanks again for having taken the time to reply with such relevant points and good ideas on how to progress.



I think the lesson with the Arranger Model that needs to be studied is not Greensill, but the USA S&L banking crisis in the late 1980’s and early 1990’s. This Arranger model seems particularly susceptible to having the ability to mint dai against assets capturable by MKR token holders and those that are able to build “governance” clout through delegation. This could be further exacerbated by what was referred to as “private entanglement” or in my mind just that once the bureaucracy is in place, it is much harder to change.

What do I mean? How would we feel if at Wells Fargo bank a large shareholder was able to exert influence in order to get the bank to make loans on favorable terms, change terms, or otherwise was allowed to source loans into the banking system. This sounds crazy, but to some degree that is what caused many Savings and Loans failures in the USA when the bankers and shareholders used their influence in the banks to make dodgy loans to their friends, families, etc. There were alot of good ole’ boy deals in these small S&L banks and really took years and a lot of taxpayer dollars to clean up.

With MakerDAO as a central bank mindset we can see how easily it would be to mint dai against sketchy collateral and I still don’t see the necessary independence of any self regulatory bodies that would prevent this from happening. In fact, I really don’t see any policy reasons that this proposed entity has to not be regulated as a meatspace financial institution. This is the reason that the “Federal Reserve”, despite its flaws, has independence to make decisions and try to accomplish its dual mandate. Similarly most banks have a separate “credit” team that is responsible for meeting its risk ratings and loan quality, and then of course these are regulated through a process of bank examiners.

To my knowledge Ethereum doesn’t have “bank examiners” or independent bodies that are charged with making sure that the capital and credit quality is there to avoid disaster. In the case where Ethereum doesn’t self regulate, I fear that regulation will be coming hard and fast. I do see the call out to form a Legal CU, but I think to the extent that the DAO insists on moving toward originating loans the first hire needs to be someone that has bank examiner experience so that they can clearly delineate the control aspects of a loan underwriting system, from ownership “token”, and manage the risk and conflicts of interests. I do not mean to claim any malice, I just have learned in lending that you need to expect that every place with a “trust” assumption can go horribly wrong and sometimes the creeping from benevolence to malevolence is almost imperceivable. The proposed roadmap for MakerDAO and Dai appear to be very tradefi. To the extent that is where it is going, it only makes sense to adopt the credit underwriting and financial oversight that banks have developed over the past few decades.

That is why I think to the extent that there is credit underwriting necessary, it would be much better to just outsource this function to existing short dated bond ETF’s. These at least come with clear credit ratings, and would likely be similar to what USDC and others are expected to do with their fiat pegged offerings that will likely be subject to “bank” examinations. It is much better to avoid private entanglement if Dai is managed similar to money market or short dated treasury ETF.

I am obviously concerned with the direction that the DAO is going and the systemic risk that would be involved if the right to mint Dai falls into the hands of a select few that can use the system to exchange dodgy loans for Dai.


Thank you for your email @jameskmccall, and thank you for the relevant points.

As I stated already in the reply to @SebVentures message, I strongly agree with you that we should enhance internal credit (and compliance) controls when it comes to a real-world function that, for the need of bootstrapping it, has been extremely centralised so far. I have been advising, investing, and participating as board member into regulated banks all my career, and I agree with you that banks do control credit risk very well - on average. Market and counterparty risk, less well, but that’s a different story. Hence, I agree with your points of pushing out as much underwriting activity as possible. A framework with an extra layer of control wants to go exactly in this direction. Such move would be, in my view, supportive of any type of credit underwriting: being it through private or through public vehicles.

Ensuring absence of conflict of interests or malicious activities is a tough exercise, but one that is helped by decentralisation and transparency in my opinion. And I truly hope that the Maker community will keep policing the activities effectively. Exactly in this spirit, the origin of my involvement was an audit of a deal that I found, as a community member and former credit professional, unsatisfactory in terms of credit risk mitigating factors and risk-reward.

For what concerns public vs. private credit, I believe this should be one of the conversations to have within a framework that we aim as agnostic as possible. Agnostic to credit types or structures. Private credit agreements have their risks, but credit ETFs are not panacea - they have significant liquidity mismatch issues and aren’t used by institutional investors at scale - those investors prefer securitised exposures or CLO-like exposures. Those same exposures I hope will be on the table when discussing about pipeline. In addition, I struggle to see the advantage in investing in money market instruments rather than in USDC which is a tokenised money market instrument itself. Growing the balance sheet through minting against short duration IG credit or money market ETFs would not be too dissimilar from expanding the dependency on USDC we have been trying to diminish lately.

I also agree with your point, which is extremely fascinating, about an absence of a regulator across Ethereum DeFi. We all hope that the system will self-regulate, but we are not there yet in any way. That is the reason we hope that real-world finance activities could piggyback reputable regulators out there: i.e. investing through regulated structures, leverage trust frameworks, collaborate with regulated entities where risk is already vetted/ rated.

As a last point, again on the Arranger Model, I would suggest the community to move beyond Rune’s initial definitions, as we mentioned several times.

I share with you the concerns over systemic and protocol specific risks, and I hope that the activities of the DAO will go in the direction of mitigating those through constructive dialogue and cautious controls.

I am eager to hear more from you on this, and I hope we can put those points into an agenda for the next 2-3 hours call.



I like these principles and believe they are (1) guiding Maker in the right direction as well as (2) inspiring a robust discussion around the key issues. Its essential that Maker develop a process that ensures high quality investments and can scale rapidly.

I fully agree that promoting DAI stability is a key priority for both Maker and RWF. With this in mind Maker should hold high quality/low risk assets that are relatively liquid. I hope we have a broad consensus here!

The Arranger will be a key key role, requiring a diverse skill set. I think it will be similar to the investment banker role in many ways but differ in that their only(?) client will be Maker. And the Arranger will need to be much better aligned with Maker than traditional investment bankers! Properly structuring the Maker-Arranger relationship will be essential!

The Arranger role will evolve over time as Luca notes but here is my view on their function:

  • Have deep expertise in the origination, underwriting, structuring and management of credit investments in asset classes where they work with Maker.
  • Be a high integrity, public interaction point between Maker and the real world
  • Ability to build relationships/source assets from good originators
  • Ensure that well aligned Originators generate quality assets which meet Maker’s eligibility criteria. This also includes requiring Originators to provide necessary asset reps and warranties and buyout/replace any assets that are later discovered to not meet these standards
  • Ensure the investment assets are serviced by strong Servicers that are motivated to provide quality servicing that efficiently monetizes the assets. Also ensuring a strong backup Servicer is in place when needed
  • Structure the overall investment so that it provides Maker with an Investment Grade risk and return in accordance with Maker’s standards. This requires not only adequate credit support but also key risk mitigants such as payment priorities, performance covenants, triggers, events of default and remedies, eligibility criteria, portfolio concentration limits, key legal opinions and other legal and tax issues
  • Obtain sufficient capital to provide required subordination for Maker’s investment, when needed
  • Oversee key transaction parties in investment including Originators, Servicers, Trustees
  • Assemble and oversee other required 3rd party servicer providers for the investment such as accountants, trustees, collateral agents…
  • Oversee all investment reporting
  • Managing investment issues on behalf of Maker as they arise, including credit issues
  • Provide timely interaction with RWF and Maker

Here are some other comments:

When I see the Arranger eligible collateral concept, it makes me think of asset backed warehouses which are usually provided by banks to asset originators. The basic idea is that the bank agrees to provide ongoing finance for newly originated assets as long as they fit the credit box, so to speak. These will include key items such as (A) eligibility requirements for each asset originated, (B) portfolio concentration limits for the overall asset pool, (C) reps/warranties for each asset originated, (D) minimum equity/subordination level for each asset, (E) performance requirements for the overall asset pool, (F) Servicing standards for the overall portfolio… Revolving securitizations are similar as well.

I suggest it is better to refer to Maker’s credit standard as being Investment Grade (IG) as opposed to an investment have less then 1% chance of defaulting over one year. Its notoriously difficult to measure an actual default proability for a given investment, but there are existing credit standards to be say a BBB bond. And these are roughly equivalent. So we can be roughly accurate as opposed to exactly wrong.

I take some comfort in commercial bank loan portfolios being examined by regulators but do not view this as a panacea. In the US I believe that the best credit risk institutions are credit investing shops such as PIMCO, Fortress… These are not regulated institutions such as banks but tend to be well managed and are run by top notch credit investors. I think they are the best credit shops because their owners/portfolio managers are top tier credit investors and are relatively well aligned with their own fund investors. They also tend to earn more $$$ than commercial bankers so they can attract top tier talent. So I dont think regulation is the holy grail for Maker

I suggest that potential Arrangers look at existing traditional credit standards for the types of assets they want to originate and see what an IG deal in that asset class requires. I expect that RWF/Maker will review the comparable traditional finance investments as a starting point. Also note that an IG deal is not only the required credit support levels, its also the broader deal structure with the types of items I mentioned in the last post. The rating agencies provide decent pre-sale/sale reports on new rated bond issues that cover many of these key issues. BUT note I am presenting my view and not a Maker consensus view!


Thank you for your message @Eumenes, and happy you are sharing the vision the report wants to communicate.

You are touching several relevant points I would like to comment on.

Liquidity I agree is crucial here, and for more than one reason. The report focuses on the effects on credit quality of asset duration, i.e. the fact that assets with lower duration/ higher churn are easier to manage and have less duration risk, but this is only part of the equation. I think that a scaling RWF exposure can significantly change the liquidity profile of Maker’s (asset side) of the balance sheet, with impacts on the PSM especially. With the use cases financed by Maker expanding, Asset Liability Management (ALM) will be more and more important up to the point where maybe a specific CU will need to be dedicated to this. We are far away from that currently, but we have learned to see the world evolving faster than we think!

The same discussion should happen if and when Maker decides to deploy the proprietary liquidity (raised through the PSM) in both on-chain and off-chain investments.

I truly cannot add on this, as I am aware the RWF is dedicating a lot of time and effort in drafting a comprehensive checklist to assess arrangers. The arranger role itself will evolve and the DD effort will have to follow such development I think.

I agree with you, and I received already a lot of comments about it. I am proposing to change the reference in the eligibility criteria to “investment grade like” type of risk rather than focusing on specific PDs. In addition, I have been thinking about it a lot, and I think that it is the Expected Loss, rather than the Probability of Default, that should be of paramount importance for Maker. With good lenders/ servicers out there able to restructure defaulting positions or to monetise first lien collaterals, Maker might find itself not in the position to do anything even if the PD is somehow higher than 1%. It is the ultimate loss that is more relevant for Maker rather than payment skipping. I will try to be more detailed in the next version of the report, but thanks for your valuable input here.

Agree. Banking regulation is a comfort point but not sufficient or necessary. In a past life I have been analyst of a 2b private credit evergreen HF and I see what you are saying here, and fully agree with you. Ultimately, we are looking for quality, seniority, credibility, and regulatory vetting of some sort. I believe that the DAO will have to be flexible and transparent in assessing each risk. In the long run, things will homogenise for sure, but again we are years away. Also, all the aspects below are not panacea I agree, otherwise it would be enough to create an algo buying public credit instruments without even looking at what’s inside and clip the yield.

I agree in full. We shouldn’t be ashamed of standing on the shoulders of giants, and improve the mechanism rather than burning everything down. I hear that your views are yours and yours only, I guess I should put a disclaimer on all my posts but that disclaimer is implicit: I am just an enthusiastic Maker community member trying to help create a vision that contribute to a more efficient, ultimately more equal, world.

Thanks again for your comments, Luca

1 Like

I greatly appreciate the diligence, effort, and thoughtfulness that @luca_pro spent researching and writing this report. To my knowledge, this is the first blueprint for a comprehensive credit risk framework within web3.0 that holistically incorporates off-chain assets—no small feat!

Luca’s two-pronged approach (identifying the types of credit Maker should underwrite and analyzing how Maker should be structured to facilitate that credit) is especially helpful in framing this discussion. There are compelling reasons for Maker to serve as a super senior lender in the financing stack, which implies an ongoing focus on accepting defined types of structured credit in a regulated manner. I fully support the report’s recommendations on these fronts.

In the spirit of brevity, I also want to highlight some of the key operational considerations that may require further detail prior to moving the proposal forward for voting.

(1) Additional Core Units

At the heart of this report is the recommendation to add a Real-World Finance (RWF) Compliance Core Unit, RWF Portfolio “Sub-Units” with specialized expertise in risk underwriting for designated verticals, and an RWF Legal Support Core Unit. Each is necessary to facilitate the structures and processes required to scale real-world assets.

An additional benefit of this design entails the creation of checks and balances between those responsible for onboarding collateral and those focused on assessing risk based on approved underwriting criteria. As others have pointed out, there are a myriad of historical reasons why banks use this structure today: different verticals within an organization can have disparate incentives driving their behavior. For the same reason, it makes a lot of sense for Maker to adopt this organizational set-up too.

While this report is understandably only the initial step towards creating these core units, several procedural questions remain to be ironed out prior to a vote. These questions include:

  • Who should lead these new units?
  • How do they integrate with the existing Real World Finance Core Unit?
  • What is the relationship between the Real World Finance Portfolio Sub-Units and the RWF Core Unit?
  • Are the RWF Portfolio Sub-Units considered “departments” of the main RWF Core Unit, or separate entities, or another specified relationship (i.e., RWF becomes an incubator that spins off RWF Portfolio Sub-Units)?
  • Will loan applications initially flow through the RWF Core Unit before distribution to the RWF Portfolio Sub-Units, or will they go directly to the appropriate RWF Portfolio Sub-Unit?
  • What is the scope of the RWF Portfolio Sub-Units’ responsibilities (i.e., do they maintain responsibility for monitoring the credit worthiness of new collateral for revolving borrowers on an existing line of credit)?
  • As we scale, how do we ensure the Compliance Core Unit does not bottleneck the process?
  • How do we adequately publicize these changes to both MKR holders and potential partner institutions, with an emphasis on the advisory nature of these Core Units (i.e., while MKR holders hopefully account for the advisory approval given by RWF and Compliance when voting, they retain the power to disagree)?
  • How do these changes impact existing borrowers? Should they be grandfathered in, or will they have to meet these new requirements or risk losing their access to credit?
  • How can this organizational structure be implemented and operations conducted in a manner that does not increase regulatory risk to the Maker protocol pursuant to applicable securities laws, including the Howey test and the Reves test?

(2) Intermediaries

The second major piece I wanted to highlight involves the importance of intermediaries in originating, structuring, and presenting loans to Maker. @Eumenes wrote two excellent responses analyzing the critical role these intermediaries will play if this framework is implemented, along with many of the important considerations to evaluate for intermediaries in this model.

In my view, the single most important factor here (aside from the core question of why the borrower needs money and what they plan to do with it) is ensuring incentive alignment through the entire lending stack, from the original borrower to the most senior lender. Intermediaries must have a requisite amount of skin in the game for any deal they bring to Maker. This prevents the “hot potato” effect, where it may be in an intermediary’s interest to originate deals just for the fees, because they can then pass the buck to the next person in line without regard for the long-term consequences of the deal. Among the many problems this scenario presents, the most notable is Maker’s positioning in the Conga Line—the next, and last, person to be left holding the hot potato. As a result, Maker has to be conservative, judicious, and sufficiently insulated in its dealings with intermediaries.

One potential solution could involve requiring intermediaries to stake a predetermined percentage of MKR tokens in an individual “borrower pool” based on the size of the deal. For example, if someone wants a $10M loan, then perhaps they must stake $1M of MKR tokens in addition to meeting other collateral requirements. This idea is at the seed stage, but I think there could be possibilities of using the MKR token in novel ways to require borrowers and intermediaries to put additional skin (5-20% of equity capital) in the game.

Overall, the securitization framework between intermediaries and Maker needs to be airtight. This will require a good deal of transparency from potential intermediaries, the public nature of which may be different from a typical lending agreement. Borrowers will likely request some flexibility from Maker on this front, but ultimately Maker must ensure the viability and integrity of potential those seeking to borrow, and the market value and liquidity of the collateral they post.

(3) Clean Money Initiatives

The last consideration involves the integration of Clean Money Initiatives for this framework. If the community decides to move forward on this front, perhaps we could create a dedicated pool of credit (Climate Credit Fund) with a hard cap (i.e., $50M to start) for climate-focused lending on an experimental basis with specific underwriting standards designed for climate-based projects.

Maker would also need to strictly define the types of climate-based projects that would be acceptable. For simplicity, only very clear-cut clean energy proposals with broad-based community support should have the capacity to access the fund initially.

Projects that meet these criteria would not be absolved from the overall underwriting process, but subject to targeted lending standards. These lending standards would then be quantified by Maker pricing in the external costs arising from the nature of the project. For instance, if a borrower sought to construct a building to a LEED-certified Gold Standard, Maker’s evaluation metrics could account for the additional value in such a project (and thus the increased cost) in ways that a traditional bank might not.

Putting the intricacies of real estate financing aside, there should be quantifiable ways to lend to climate-focused projects using expanded underwriting criteria that account for the long-term social and economic benefits of investing in this sector, the potential for increased value of climate-based investments—especially as costs that are currently externalized become internalized (imagine if a price gets put on carbon in major international markets—investment calculations would rapidly change), and the climate financing market share Maker could take as part of this push.

The primary way to balance these risks would be to start small with the aforementioned hard cap. The community could then set verifiable benchmarks and check the Climate Credit Fund against them at set intervals. If those benchmarks are met, we can linearly increase the size of the fund. If not, we need to iterate and fix the process.

We’re excited to see this process continue and look forward to continuing to work with the community as these recommendations progress towards concrete proposals!