The Arranger Model - a highly scalable RWA framework

The Arranger Model

Real World Assets are one of the biggest opportunities and challenges for Maker, and the thing that sets the Dai stablecoin apart from any other DeFi project.

Just as it has incredible potential, the field of RWA is also extremely complex and difficult. For this reason it is extremely important that MKR holders structure the entire framework to make it as easy as possible for them to manage its risks, especially at scale.

This post is intended as a simplified overview of the Arranger Model. It has been written with input from some of the RWA experts in the community, but it is just a starting point that will need to be further discussed and refined publicly by the community. Luca Prosperi is currently working on a very detailed report that gives a more formal and professional opinion on how to structure highly scalable RWA capital allocation, which will help us to further refine the frameworks.

The two major risks of scaling RWA

RWA complexity and non-standardization

The biggest risk of all is the inability for MKR holders to understand the risks of the RWA collateral that is onboarded at large scale. Traditionally, the model Maker has followed with collateral, including real world assets, is a passive system where the community waits for projects to come with proposals of any kind, and then considers each of the proposals individually.

This is bad for two reasons, first of all it means that Maker is supposed to be able to price the risk of literally any type of deal that could possibly exist, which is not realistic and leads to Maker doing business with a lack of knowledge.

The second problem is that because the process is so broad, and as a result extremely slow and uncertain, we are self-selecting to only engage with counterparties that have no option to go anywhere else, and as a result we end up with generally bad deals on top of the fact that we are badly equipped to analyze them.

There’s a saying that if a borrower is willing to sit around and wait for the deal to go through, then it’s never a deal you want - because you are literally the only person on the planet that’s willing to accept those terms (or they would have gone somewhere else).

The solution:

The solution to this problem is to adopt a model based on Eligibility Criteria, where Maker proactively specifies what types of proposals it’s looking for, and creates an expectation that there is a high likelihood of acceptance if proposals fulfill the requirements.

Instead of waiting for people to come to us with completely new things in structures we are not familiar with, we should standardize and specify what we are interested in as much as possible, and then create a high likelihood of a quick and successful process for the collateral that meet our requirements both technically and in spirit.

Standardization of the content of the deals are crucial, but standardization of the structure and enforcement is just as crucial. There is no reason why we should deviate from anything other than the strongest possible legal setup that provides us the strongest possible recourse, and there’s no reason why we should spread our attention across many different idiosyncratic models instead of focusing and specializing in one specific structure for each jurisdiction.

Private entanglement of Mandated Actors

The other major risk of RWA scaling comes from the extremely complex relationships and interactions that it takes to set up a fully functional RWA pipeline with legal structures, deal flow, high value counterparties etc.

A big part of the process of performing due diligence on an RWA counterparty involves reviewing confidential information.

Without proper separation of powers, this could create a risk of private entanglement where the DAO is effectively unable to maintain relationships without the mandated actors that control them.

Examples of failure modes include:

  • Not being able to monitor the risk of a deal without the consent and involvement of a privately entangled mandated actor.
  • Not being able to liquidate or restructure a deal without the consent and involvement of a privately entangled mandated actor.
  • Not being able to access legal information of structures that are supposed to benefit the DAO if a privately entangled mandated actor controls the relationship with the lawfirm

Because it’s hidden as it builds up and only emerges after the damage has already been done, private entanglement can cause tremendous buildup of hidden risk in the long run if it is not properly controlled.

The solution:

The way to control private entanglement of mandated actors involved in RWA deals is to strictly enforce separation of powers and decentralization across the entire RWA pipeline. There needs to be strict requirements around how mandated actors can engage directly in deals, and generally their personal involvement needs to be minimized. Mandated actors must facilitate interactions between counterparties and the DAO, rather than act on behalf of the DAO and personally interact with counterparties, to the extent this is feasible.

There also needs to be internal controls specific to RWA, including auditing, standardization of best practices and enforcement of transparency standards. There also must eventually be multiple redundant RWA Core Units, to prevent a single point of failure from emerging, even at large scale.

The arranger model

My proposal is to create a standardized framework based around the concept of “Arrangers” - professional and highly experienced financiers with extensive networks that get pre-approved by Maker Governance based on recommendations by RWA Risk Core Units to source, arrange and propose RWA deals with Maker that fit the Eligibility Criteria Maker Governance has proactively adopted based on committee discussions and recommendations from the RWA Risk Core Units.

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 terms of the Boxes are then continuously changed over time based on the performance of the Arranger, mostly by increasing the debt ceiling over time.

Examples of Boxes would be:

  • Renewable energy projects in the US

  • credit tenant leases in the US and Canada

  • Short term institutional debt in x region

  • etc.

Each of these boxes have their own terms, including equity requirements, maximum and current debt ceiling and interest rates. Technical DC control could be implemented as a single vault with a very slow version of a DC-IAM, while the Boxes and their individuals terms aren’t implemented on the protocol side but instead through the legal structure.

Around the arrangers is a self-regulating system that supports massive scaling, and implements a strict separation of powers with many distinct categories of actors that work together to implement the scalable and secure RWA funnel.

RWA Audit Core Unit

The RWA Audit Core Unit is a secondary (meta-governance) Core Unit consisting of a relatively small team of Highly experienced credit professionals. e.g. credit risk management experience in all stages of the credit process (origination, modelling, underwriting, performance monitoring, collections and recoveries)

The RWA Audit Core Unit continuously works to create and evolve a ruleset for best practices in terms of legal structuring, transparency, decentralization in how the RWA Risk Core Units conduct themselves. The RWA Audit Core Unit then continuously monitors and reports to Maker Governance about the performance of the core units, and if it detects potentially problematic behaviour it will attempt direct remediation efforts and ultimately recommend action by governance if it is unsuccessful.

RWA Risk Core Units

RWA Risk Core Units are primary (doing-the-work) Core Units that do risk analysis and make governance recommendations related to the onboarding of Arrangers and specification of their Boxes, modification of Box terms for Arrangers, and in exceptional circumstances, the liquidation/unwinding of Arrangers. At scale the RWA Risk Core Units themselves are meant to be internally separated, meaning rather than building up large internal teams in black boxes, they have lean mandated actor teams that administer the Core Unit and its privileges such as budget an executive vote privileges, while managing a decentralized ecosystem of private actors doing professional risk services to perform risk analysis on the Arrangers.

Onboarding of Arrangers

RWA Risk Core Units in practice engage in discussions with prospective Arrangers, and based on these discussions and the analysis of their risk microservices resources, they recommend one or more boxes with appropriate terms.

Modification of Box terms for Arrangers

There are different conditions that can trigger an analysis that may lead to a recommendation to modify terms of an Arranger Box. If an arranger has hit the maximum debt ceiling of a Box, then it can be considered to increase it even further if the conditions and the performance allows for it. It may also be that an arranger wishes to specialize in other areas, or that they have performed inadequately and need to have their debt ceilings or terms made more strict.

The RWA Risk Core Unit makes the analysis, potentially relying on appropriate risk microservices, and then if they think it is necessary, makes a recommendation to Maker Governance.

Liquidation of Arrangers

In some special circumstances conditions may occur that make the RWA Risk Core Unit analyze whether it is appropriate to recommend a liquidation of an arranger.

Arrangers

The Arrangers are the key actor in the RWA pipeline, as they are the hubs of all the different commercial and legal relationships that it takes to set up a solid, scalable RWA structure. They typically specialize in a few specific areas, such as specific regions or type of credit, and then focus on scaling it up as much as possible within the autonomy they are given by Maker Governance, based on the recommendations of RWA Risk Core Units.

Arrangers are highly scalable business models that focus on charging small margins on quantity, and there is an economic limit, as well as a “political limit” to the amount of Arrangers that can practically exist, since their business models have to reach enough scale to be viable, and they need to be politically present in the community so they can build up enough trust to get the necessary scale.

Arrangers can set up various pipelines and structures below them, but in all cases they have to fit into the Boxes that have been approved by Governance. This gives them a lot of flexibility, and most importantly gives them the ability to proactively seek out deals and execute them without having to wait around for the bureaucracy of the Maker Governance process to sign off on them.

In practice however, they always need to consider the will of Maker Governance, since even if individual deals don’t need direct governance signoff, if they deviate from the spirit of the deal, e.g. by stretching the definition of sustainability or getting around sustainability metrics based on a technicality, then it may result in them getting their Boxes modified to prevent them from expanding any further. This mean the best interest of Arrangers is to operate as much as possible in alignment with the will of Maker Governance, but with the benefit of having no red tape or bureaucratic roadblocks on a day to day basis, and with the autonomy to make decisions on individuals deals so that they can be closed in a way similar to what the traditional financial world is used to, making Maker competitive as a lender.

Attestors

To minimize overhead and complexity of the RWA Risk Core Units, the leg-work of analyzing data and continuously monitoring the Arrangers is done by Attestors, private actors that work with multiple Arrangers simultaneously. Arrangers are pre-approved and by Maker Governance in the same way as Arrangers, and Maker can put limits in place to ensure that there are multiple Attestors across the different Arrangers, minimizing single point of failure risk.

Attestors, as private actors, also have the ability to enter into NDAs and analyze proprietary data, and then publicly attest to its validity.

An Arranger needs to already have an agreement in place with an existing, pre-approved Attestor before they can be onboarded and get Arranger Boxes and a debt ceiling.

Trustees

A key element of the Arranger Model is to standardize the enforcement of claims against the collateral, by only working with top-tier trustees, and ensuring that no cash flows can “leak” outside the control of the trust.

Maker must pre approve the list of trustees it considers top-tier for each jurisdiction, and arrangers can only be onboarded if they already have a relationship with at least one such trustee.

Lenders

The Lender is the entity that borrows money from the Trust, and then allocates it to the various Boxes that have been approved, by lending out the money to end-borrowers.

All capital allocation must always be in senior secured credit.

Inherently Trusted counterparties

Some potential RWA counterparties to Maker are entities that themselves are as reputable, regulated and capitalized as top-tier trustees. Examples would be large regulated banks or other large, regulated financial institutions.

Inherently Trusted counterparties can make proposals directly to RWA Risk CUs without the need for a pre-approved Arranger Box, and RWA Risk CUs can directly recommend onboarding of their vaults, and modification of their risk parameters to Maker Governance.

An example of this would be the recent Societe Generale proposal.

Potential Arrangers ready to start scaling

In the community today there are already at least 3 groups that have potential to act as highly scalable Arrangers in this framework, and scale to significant size, potentially several billions, within a year.

Matt (6s Capital) is the first RWA provider to implement a trust based structure. They are US focused. The experience of Matt and strong structure of the RWA setup means that rapidly scaling to 100s of millions in the short run is realistic.

Christian (MD Radiance) is a veteran from the RWA Core Unit that is now potentially ready to onboard as an Arranger. MD Radiance has indicated to me privately that they are interested in taking over the live and near-live Centrifuge deals, so that they can smoothly be restructured to work under the Arranger model, without disrupting our partners access to credit our damaging our credibility. In addition MD Radiance is able to focus on Canada and other non-US markets, which make them well positioned to handle Backbone collateral at very large scale (potentially 1 billion or more within a year).

Allan (Monetalis) is a newcomer to the community but with significant industry experience and based on what I’ve learned has the interest and the necessary skillset to act as an arranger. Monetalis is UK focused, which makes it a great place to place Backbone collateral that can diversify our exposure out of the US. They will need strict analysis but if passed my estimate is that this is another potential of up to a billion of non-US RWA collateral within a 1 year time frame.

In total, we already have everything in place that allows us to diversify up to 2 billion USD worth of collateral outside the US, and do major flagship projects inside the US also potentially at the scale of billions.

Since 6s has already proven the basic pipeline is legally viable and works end to end in a live setting, it is going to be a matter of Maker Governance creating a stable enough environment to give these businesses the confidence they need to start scaling up their operations and fully engaging their networks.

Meanwhile we should also consider onboarding more Arrangers, especially outside the US in climate resilient areas such as New Zealand.

Maybe even consider increasing the DSR to preempt the situation where we have too much collateral and Dai demand becomes the constraining factor on growth.

20 Likes

Thank you for this @rune, I am excited to be part of this journey.

As @rune mentioned, we have been working on detailing and formalising a new way to institutionalise and scale our real-world finance efforts in a project that we called Real-World Sandbox. I have been mandated by the SES Core Unit (special thanks to @Juan and @wouter for their support) to lead the project, and have received tons of inputs from the community in private. All those inputs are currently in the process of being incorporated in a report we will present in the next week.

We also have a forum thread where everybody will be able to participate publicly to this, where you can also find my contact details that I paste here below.

email: [email protected]
forum handle: luca_pro
Discord handle: luca_pro#7418

Also, for those who are curious about my views in general, I proudly spend my nights publishing a weekly newsletter that I have called Dirt Roads, you might have encountered it already.

As always happens, the ultimate process the project will propose might differ in the implementation details from what @rune has been proposing above, but we all share the same ambition of making of Maker a reliable, institutional, conservative, and scalable senior financier of the most impactful and transformational projects happening in the real world.

I personally believe this is an incredible time for DeFi. A time when the infrastructural and philosophical visions that started it all can kick-off a virtuous feedback loop that could impact a much broader community of humans. The Maker community is the best positioned to become the pillar of this next phase.

I am honoured to participate to this movement.

Luca

12 Likes

Interesting. WIth family for holidays and have other things on plate here.

One key point not included in the above is ‘conflict of interest’. I have repeatedly seen that MKR holders are also investors in other projects (i.e. Centrifuge as one example), or have interests. One of the biggest issues for legislation is for ‘disclosure’ within business relationships of ‘significant ownership interest’ and ‘conflicts of interest’.

Where does this fit in your model.?

Second point as a physicist who has worked in many fields what you write with your arranger model is similar to a term called.

‘acceptance’ in beam physics. What this means is if the input of the product is within the acceptance space of the production unit it will pass through processed. If it is outside of the acceptance it will end up appearing in the processing loss profiles.

I think being clear in terms of what the Maker acceptance criterion is will help line up the RWF processing so we reduce losses. Which begs a question in your Arranger model do you anticipate structuring an acceptance that is highly broad or narrow? The reasoning here is if the acceptance is designed to be broad then we can theoretically process all kinds of offers, but if it is narrow we will have to be highly restrictive. Generally in production science there are various reasons to design the broadest acceptance possible because over time the acceptance tends to narrow due to ‘machine performance degradation’. Here there might be a reason to be clear about where MakerDAO is going to narrow the acceptance so that whatever is presented is processed with high speed and efficiency, with very low loss.

I wonder if you have thoughts on this subject? This is particularly true when it comes to jurisdiction and potential for capital deployment through scale and legislation? The US has huge capital but terrible jurisdiction to navigate, where as countries in Africa may have less legislation but not much capital (even though if we could get there the actual environmental impact may be greater).

5 Likes

Just to shed a light after 1 year of actually doing RWA.

Small asset originators’ deals don’t scale. That was known from the start but was the composition of the MIP6 flow initially. This is why, asap, we moved to asset managers (1 issuer managing a diversified pool of asset originators). FortunaFi is an example.

Then you have a single point of failure which is the asset manager, which, usually, is a young small company so risky. Doesn’t make any sense for scaling. Moreover, MakerDAO become usually the main/only client, therefore you shut down, you kill them. They also usually don’t have the cash to put skin in the game. And you keep the private entanglement and subjectivity. Just by naming them and using your influence, you are putting insane pressure on whoever will evaluate that from risk.

The logical conclusion is therefore to go directly on off-the-shelf products or institutional asset managers. This removes all conflict of interest and avoids creating a bureaucratic structure, therefore keeping MakerDAO lean.

The Arranger model (or what we called non-institutional Asset Manager) can be kept for tactical allocation (when no one else can provide such product and for something the community really wants) or strategic (innovation) but should be considered riskier. We decided to propose to keep those under the Surplus Buffer to avoid tail risks. This was also part of the reasons of a limited number of slots. Not saying it’s a distraction but not that far when you are able to tap into institutional offerings.

That was the idea behind the ETF investing and a high reputation asset manager was about to put a MIP6.

Don’t you think that using institutional investments are better than your model both in term of RWA complexity and private entanglement (extending it beyond mandated actors by the play of influence)? If not why?

In addition, could we have the list of the RWA experts?

10 Likes

I think the US capital base indicates that its legislation supports its scale. While it may be complicated, subject to special interests, loop holes, etc., after having spent many years working in Africa, “less legislation” is actually a major problem. The “rule of law” is a critically important construct for RWF.

An interesting side observation to the “Clean Money” proposal is that each of the “Super Countries” have strong “rule of law” characteristics - English common law for 3 of the 4.

You can finance assets in Africa for sure, but the “political” aspect of the financing (and enforcement) becomes very important. As a consequence, lenders like to have collateral that they can enforce offshore.

Having said the above, I fully agree that climate impacts Africa very hard and the potential climate mitigations in Africa and other non-OECD countries need to be very creative. It becomes imperative to create robust structures that can both address the actual environment and the legal environment.

6 Likes

What sort of timeline do you envision for onboarding those $2B?

In general I like the energy of yours and some ideas are very valid but however many the superpowers Matt or Allan might have it’s one to assign them the capability to deliver a billion of deals each but execution is paramount here… Each actual deal takes a few months to discuss (larger the deal, the longer it takes), to act on it, to deliver it… when it comes to Real Estate projects then there’s the permits, actual construction… It is sooooo complex.

I know many highly skilled individuals across EU and US who built up investment companies (often times having previous A-class IM/AM experience) and it took them years if not decades to pass the 1B mark… Unless you just want to hit the number no matter the quality of the projects onboarded…

6 Likes

Looks like some ‘pitching’ is allowed (@rune @luca_pro)? as a potential Arranger/non-institutional Asset Manager who is going through Untangled MIP6 onboarding risk assessment, let us unashamedly put forward some of our ‘value props’:

But first, like many other applicants so far, we are just a start up - so, by definition, are limited in our resources including the so called ‘skin in the game’ (but we recognise the importance of skin in the game and generally we try harder).

For the above, potentially, minus point here are some potential pluses we bring:

Trade and green finance asset classes/Boxes: we are into trade finance and green assets which are still outside of institutional asset manager’s focus (to @SebVentures’s point). Today you can’t buy a trade finance or a green asset ETF.

Emerging markets focus: our focus is emerging markets (Asia, Africa, LATAM but some in Europe too) - so here is the opportunity for yields, high and equitable growth around the world - e.g. financing a rice exporter from Vietnam.

Ability to scale: we could have asked for a higher initial DC but chose to have a 10m DC, not because we can’t scale, but because we recognise, like @Tru_Greg, it takes time initially. We are onboarding 3 asset originators/lenders with our fist fund but are in various stages with 15+ others. Taken together they could bring 200m RWA to Maker, in the next 12-18 months.

Our RWAs are on-chain - we have built a tech platform that tokenises RWAs at scale. It allows Maker and others to monitor the asset status on-chain with trigger reporting etc. Execution is key as @MakerMan mentioned and our platform (combined with experience, ability to source quality lenders, and standardised structure) enables us to scale fast.

We use standardised, best-practice legal structures - securitisation, repackaging transactions are not new and, talking to our originators, there are well-recognised structures, documentation and jurisdictions (this, for an asset class like trade finance, gets over the issue @christiancdpetersen mentioned i.e. the lack of rule of law in Africa). Here is a source to share with the community Standard Repackaging Documentation. The remaining intricacy is to do with the fact that the DAO can’t transact directly and some regulatory headwinds with crypto/stable coins. But there are projects like 6s solving this.

Last, but certainly not least, we help create a demand for DAI (as opposed just supply - to @rune’s point) given trade finance are cross borders - we would offer fund lenders in DAI so that DAI could be eventually be used as a trade currency (now frankly either just fiat USD or perhaps soon USDC)

Do we have other options (to @rune’s point about not wanting a borrower who sits around…)? Yes, we think so - given what we have built. We are still around after @rune’s first visit to our slack to comment on a potential project in Africa over 3 years ago). We hope we bring justifiable values - diversification, scope, scale, experience, transparency and demand for DAI - to the Maker’s community.

3 Likes

There is much to like in the Arranger Model as it does highlight many of the current risks in the current RWA strategy. My concern when the DAO began its foray into real world assets is exactly the corner that I think we are painted into. A free and decentralized token then needs to pivot to “boots” on the ground and jurisdictional hooks, and human picks and shovels to make any of it work. The plan, to have open MIP6 proposals for people most familiar with token investing to make decisions on credit from paper proposals is obviously fraught with risk. It was this exact criticism that launched me into trying to protect the DAO from these risks, while still trying to find the sweet spot of the deal in the middle as the DAO had signaled that it was interested in pursuing this path despite the difficult path.

The current model does turn into a credit spaghetti test to see what sticks to the wall and the highlighting of “Private Entanglement” or I think what I would call that lack of privity of primary parties to the DAO is also an important concern. Having a “box” of credit parameters and investment allocation is probably a good idea.

I still think that philosophically, I still don’t get the strategy. In my mind there are bond guys and loan guys. These models are completely different and require completely different staffing and gearing to make work.

Bond guys - they love buying and trading liquid debt instruments and think in terms of bips and concavity and quant math. They love being able to carve a profit off of a trade on interest rate exposure or macro trends. These traditional finance types love S&P rated bonds and love nothing more to speculate on the Fed to determine whether that makes bond go up or bond go down. These models can be seen in PCV projects where they want to have a stablecoin by creating the equivalent of a money market or Treasury ETF. The DAO has many of these voices, IMO.

Loan guys, credit - This is the model that I live in as a loan originator model. I guess a “Arranger” of sorts. This model is right smack in the middle of the borrower, the lender, credit, legal, title companies all trying to kill the deal the moment that it gets brought to the “Arranger” or even in my world the “Company Man”, the field representative. The dual model is that the loan originator would try to bring deal flow to the lender, the field representative puts eyes on the collateral and also has own knowledge to apply to loan originator to provide checks and balances and due diligence form perspective of the “Company”. The title companies make sure that there is provenance as to ownership and marketable title, legal is always a battle as they basically get paid to try and kill deals, credit is the same to some extent (and both are very important to the process). My overarching point of this is to highlight that it isn’t enough to just say, I am going to protect the DAO or the “Company” as the case might be. For the loan model to work, the parties also need to make sure that the borrower is going to win as well. The Arranger is right in the middle having to make sure that everyone wins and nobody loses.

All of this goes to say that I don’t see a material difference between this model and that of a traditional investment arm you might see in an insurance company. What are the competitive advantages of doing it this way? I guess we can see. My concern is one brought up by @MakerMan regarding conflicts and fiduciaries.

The DAO is token governed, and yet doesn’t seem to be willing to follow the token based governance procedures. In a loan model, when there is a “commitment” it means something. It has to or else the borrower is going to get wrecked. I can’t really think of a really good analogy for this, except for that there doesn’t appear to be any state memory of the MIPs, signals passed, and debates had and battles overcome. There is no equivalent of “stare decisis”, that the thing has been duly debated and passed. The power center of the token votes can effectively line item veto any decision at any time.

Which, I have always been a proponent of the golden rule, “that he/she with the gold makes the rules”. My issue is that I am not sure how one is willing and able to “trust” that the shifting power centers can’t switch the direction like a weather vane. How does a loan originator build on this stack without fearing that the rug gets arbitrarily pulled.

5 Likes

There’s much to like about this proposal: the drive towards decentralisation of RWAs, a structure that intends to scale, the ability to focus on quality collateral that has a positive impact on the world.

As someone who is both a Maker supporter and MKR holder, as well as someone who has the right skills and team in place to be an Arranger, I have queries on both sides:

  1. As a MKR holder: This proposal broadly assumes that the RWA team has found a product/market fit that they want to scale. From following the RWA journey, I’m not sure this is the case yet? Please correct me if I’m wrong.
    If the community believe they have not yet found the right product/market fit, then more time needs to be spent figuring this out (legal structure, onboarding, risk assessment etc) through experimentation and ultimately trial/error. Having built real estate companies before, they are not like tech businesses that can easily ‘pivot’ when needs be. Trying to scale before all the nuts and bolts are in place will be a disaster.

  2. As a potential Arranger: I am beholden to one capital source forever. My business lives and dies off Maker governance decisions. That’s fine if I’m economically aligned with Maker (eg Maker have bought x% of my company for y MKR tokens, hence I want the best thing for Maker as a whole as it grows my balance sheet), but I don’t think that’s being proposed here. So my business lives and dies at Maker’s whim. No entrepreneur wants to run their business like this. This is closer to an employee relationship.

A solution to point 1 could be that the first few Arrangers exist with the understanding that they are decentralised members of the network who are running a series of tests in the RWA market in partnership with and on behalf of the RWA team. This allows the community to learn more, faster, while still having full oversight and understanding of the risk being undertaken across the capital stack (this learning won’t apply in going direct to an institution, eg Soc Gen). One Arranger could run 3x different strategies - currently to get that same amount of learning the RWA team would need to onboard 3x different originators.
This also helps point 2, as the Arranger is working with Maker in the first instance to test strategies, as true partners, giving each party time to figure out the best route forward. If the parties figure this out together, then long term success is much more likely. This makes the Arranger less wary of risks around Maker-dependence (they’re figured out a joint strategy so why end the relationship?), and gives the Maker community more comfort that a truly stress-test strategy for RWA is being scaled.

I like the Arranger model concept, it’s a step in the right direction of decentralisation, but there needs to be a way to align objectives between the Arranger as a Company and Maker as funder of the company (albeit currently lender on assets, not shareholder). Maker is bringing all the capital to the table, which should be recognised, but the Arranger’s need to exist outside of Maker’s financing also needs to be acknowledged. Or else you just end up with a bunch of salaried hires running an Arranger team, bankrolled by Maker, rather than the experienced and ambitious entrepreneurs who want to scale their pipeline and make the world a better place with an innovative funding partner (Maker).

6 Likes

Arranger Model - Thoughts

I have set up/restructured asset management functions for insurers and pension funds previously (as part of private equity buy-outs) - and conducted due diligence on many. With that background I have tried to map the Arranger model onto what I’d consider somewhat ‘common practice’ (not necessarily best practice, but workable practice) in setting up the asset management function. My hope is that this mapping might spur some more ideas about practical implementation of the Arranger type model in the community.

When I mapped it out I got to this picture:

Not too complex - so just a quick run through of it here:

A/ Investment Objective

1/ Everything starts at Investment Objective - somewhat the overall strategy of what is expected to be achieved in both qualitative as well as quantitative measures. I have seem good examples where they at least addressed these items:

1.1/ Investment Principles. For instance:

  • We only invest in ESG assets
  • We focus only on credit or credit-like instruments
  • All managers we work with must provide a diversified portfolio
  • We only accept senior instruments
  • etc.

1.2/ Cash-flow and liquidity Demands (from the liability side). For instance:

  • Portfolio must deliver xx% in actual return/yield cash flow every 3 months (to fund operational costs)
  • xxB USD must be able to be liquidated within 30 days and yyM within 24 hours
  • etc.

1.3/ Risk Tolerance. For instance:

  • E(L) for any portfolio must be <1%
  • Overall portfolio must have certain CVaR and Sharpe properties
  • etc.

2/ This, generally, is developed by some key units who know both the asset and liability side and general strategic direction of both. I don’t know exactly the scope of the various CU’s in Maker, so I put in ALM CU (the asset liability management CU - basically the CU that would have the responsibility for A&L side matching whatever that might be today) and the Portfolio CU (the unit responsible for asset allocation under the general investment policy - whatever that unit might be considered today). Usually - in practice - the responsible units would seek inputs from the audit and the risk units etc. - before coming up with a recommendation to ‘the Board’/MKR holders.

3/ The Investment Objective would then be discussed and reworked until ‘the board’/MKR holders approve it - usually would require some supermajority or such to clear an Investment Objective.

B/ Investment Policy

1/ With that in hand, the Portfolio CU and the Audit CU would get busy and need to develop a practical asset allocation and monitoring plan. I have seen them being quite detailed. Involving fairly deep views on how audit and compliance must be undertaken in the various asset classes - and substantive simulations to max return under min risk exposure with hedging scenarios etc.

2/ It should cascade from the Investment objective and get specific about a fairly large number of items - this is a rough, incomplete, list:

  • Roles, responsibilities and delegation (approval, oversight etc)
  • Asset Allocation into classes (and documentation on each class scope/scale)
  • Benchmark set for each class (return, fees, legal structure etc)
  • Due diligence requirements for onboarding mandates/AM/Arrangers
  • Monitoring and reporting requirements per class/mandate
  • Risk Management design and requirements
  • Custody requirements
  • Liquidity mechanism
  • Hedging requirements
  • Leverage requirements
  • Other/Admin: Accounting for fees/expenses/interest, Valuation of Assets, Rebalancing mechanics

3/ Normally - again - the responsible unit would seek input from the risk units, the AM/arrangers, the attestors etc to paint this comprehensive picture.

4/ As with Investment Objective, the Investment Policy would need approval by ‘the board’/MKR holders.

5/ Why not do Investment Objective and Policy in one go? Objective should have more longevity and be more principle-based than the Policy - and it seems better in practice to anchor discussions around two different time horizons.

6/ To a large degree, in the longer term, the asset allocation is a larger determinant of return and risk to the overall portfolio than the selection of the individual asset managers/arrangers (given prudent selection, it is hard in the long-term to escape asset class average performance for any individual asset manager). The importance of this step of deliberate allocation is quite important in my experience for overall performance of the portfolio.

C/ Asset Class Allocation

1/ So usually, ‘the board’, steps out of the way and allows the Portfolio CU and Audit CU to execute - expecting the reporting and monitoring system put in place allows for ‘the board’ to be alerted when something is not right and/or there is a need for rethinking the Investment Policy. This, of course, might not be the case in a DAO, but the key is that focus now is on effective execution rather than debate on principles or approaches etc.

2/ So, usually, you’d have various asset class specific units with appropriate experiences, who would now be given a mandate based upon the Investment Policy. I am thinking those are sort of like the envisioned Risk CU’s I reckon. If a CU was needed according to the Investment Policy, it would be incubated/set up. For example (just examples!) one could have Risk CU’s for:

  • Renewable Energy Project Financing
  • Forestry & Agriculture
  • SME lending
  • Corporate Bonds
  • Invoice/trade Financing
  • Etc etc

3/ Just to be clear, these units do not have to be big: I have seen 1 or 2 man teams lift many many USD 100M’s allocations and doing so very well.

4/ Now the portfolio CU (or whatever unit has this role in Maker today) would continue to play a large role in aggregating and monitoring across the Risk CU’s in terms of monitoring, reporting - and often also being responsible for ensuring cross portfolio hedging (and even leverage) is arranged - and accounting and risk management measuring done. (for this reason sometimes this unit is split into subunits: hedging, accounting, risk mgmt, portfolio mgmt and simulation/analytics etc).

5/ This is also the time where the Audit, Compliance and Practice unit earns its stripes. It now needs to provide the necessary monitoring and audit for the portfolio unit, the risk units, and via the Attestors (sort of like external auditors and accountants I assume with a Maker specific charge or such) to ensure they are in compliance with the investment policy decisions and investment objective principles.

6/ I wrote Practice into the Audit CU, because I get a sense that there needs to be a guardian for a number of standard operating mechanics/practices that must be in place irrespective of asset class. As, for instance, in how the various asset managers/Arrangers are legally tied into Maker vaults, which trustee’s are ok, which banking relationships are above board etc…

7/ There usually would be some natural (and desired) tension between the audit/compliance and the risk units (as I have described them here anyhow!). In the simplest of terms one is the gas pedal and the other is the brakes - together making sure that, as we drive down the road laid out by the investment policy, we don’t run off it, but still get somewhere.

8/ There would usually be, at the very least, an accounting department and some legal department to support this structure. I am to a certain degree implicitly putting the legal department into the audit CU and the accounting into the Portfolio CU. Could easily be a different setup of course.

9/ The Risk CU’s now need to fill-up their dance cards: finding Arrangers and AM’s and whatever instrument that is supported under their mandate from the investment policy - onboarding them and monitoring them etc. Nothing new there.

10/ It wouldn’t be uncommon for an AM/Arranger to receive more than one mandate from 2 different departments in a fund - for instance a pension fund might give, say, Morgan Stanley, the investment mandate for a bond strategy in the US from the bond team and a growth equity mandate from the equity team in UK - and the Green renewable energy project mandate from the ESG team in Switzerland. I suspect that sort of cross-mandates would potentially happen here also?

11/ As you may recall, at this level of execution ‘the board’ is rarely involved (but could be different in a DAO I understand), but often there is an Investment Committee (with only internal participation - or sometimes with intermixed relevant external advisors), that approves each onboarding of an external asset manager/arranger.

D/ Investment Mandates

1/ We’re at the end now - Asset Managers/Arrangers do what they are supposed to do under the sub-mandate they have received from the Risk CU’s and allow Audit CU/Attestors in for audit/monitoring as required.

Observations
This ‘mapping’ lead me to a few observations - things perhaps to think about:

1/ I don’t quite know what core units today have the roles of ALM CU/Portfolio CU I mention in there. Maybe some of it sits with RWF today? At any rate important, in my view, that these roles are clearly assigned somewhere. That will help, I think, in driving a good asset management strategy. .

2/ Even if people today might think they feel quite clear on what the Investment Objective and Policy are for Maker (personally I am actually not entirely sure everybody thinks exactly the same about it), I do think it makes sense to formalise and institutionalise it. Helps to keep aim straight and somewhat disconnects it from persons and interests. Also makes implementation and execution efficient - and scalable. Even if the resulting strategy is as simple as ‘only buy t-bills and hold’, then codifying all the elements above usually will make sense in the name of efficiency and institutionalisation.

3/ Should Maker adopt a process somewhat closer to the one sketched broadly out above, Maker becomes an easier animal to work with for professional asset managers and institutions, I’d suggest. They have seen this process before, understand it and can easily work with it.

4/ In practice, making use of “expert best guess” (i.e. you let a few experts build the investment objective and policy VERY fast based upon their experience and then you edit in consultation with the community and get MKR holder vote etc) Maker could probably get this process kick started very fast and have a working frame a la above within quite a short time-period - and then in parallel have a longer process going to get it exactly right.

5/ Importantly, if well structured, this type of process/framework makes it such that you do not need a lot of people to execute - as long as you just get high-quality expertise into the roles you can go with fewer than one might think. There is this measure I like to keep an eye on: “Cost of Governance”/AuM - i.e. the cost of running this sort of process divided by assets under management by the structure. This metric should be in a forever downwards spiral - or something usually, in my view, is wrong with the implementation.

6/ Hard to discuss asset management without discussing compensation. The short and long of these conversations have a tendency to come out like this in the end:

  • Audit team needs comp not connected to any investment performance (to be independent) - also for Attestors I suppose if I understand the role right. Great audit/compliance people are expensive (and increasingly so) - but I have also found it never worth trying to save on this element in practice.
  • Portfolio CU + Risk CU (as I have scoped them!) needs comp connected to overall portfolio performance in allocating to fit mandates (in a few dimensions) and benchmark return (to understand they are part of a whole - better than only connected to their allocated asset class performance - encourages talking to each other).

7/ Another common discussion point is building internal capabilities to ‘invest direct’ vs. using external asset managers/arrangers. I have seen pro’s and con’s for both. Harvard Endowment recently dismantled a mixed model to work only with external asset managers, because they didn’t feel they had the scale in the individual asset classes to build enough skills - the total endowment is about USD 50B. On the other hand I have seen pension funds establish excellent internal capabilities with great return on key special asset classes (energy infrastructure, warehouses, private debt etc). I guess the key here is to go in with eyes open and consider whether Maker can in-house enough skills and expertise to compare with what can be ‘purchased’ from an external party.

This was a bit of a “mind-dump”, but I do hope a few of the thoughts might bring out some good ideas from community members on how to practically get an Arranger type model designed and implemented.

I will, for sure, keep thinking some more about it.

9 Likes

Thanks everyone for the amazing replies. I think threads like this where every single post adds knowledge and insight are incredible and show how crazy the potential of the DAO is.

Took me a while but here’s a wall of text to respond to some of the great input

Just a starting point for discussion

Most importantly I want to stress that the original post is not meant as a finished, “authoritative” framework that’s ready to be implemented and used as is, but rather a simplified starting point of a potential direction for RWA that is centered on the concept of “arrangers” as the intersection point between the real world and maker governance, and the structures that they put in place are “set-and-forget” meaning the arranger has no leverage or trust assumptions once the collateral has been approved.

I think in particular the upcoming RWA Sandbox report by Luca and others will have a bigger impact in setting out something that is more relevant to actual implementation, and also comes from expert sources.

Institutional products

Personally I agree with Sebs point that we over time will get most of our scale through institutional products because that’s the only place the capacity exist to deliver quality at scale. Others including Luca have made that argument as well so I think we can count on it. It won’t take long until all the big institutions will be knocking on our door if we can demonstrate a solid process and low cost of capital.

The arranger model could still play an important role in solutions for onboarding such assets, because the arranger in the most general sense describes the point of interaction between governance and the real world assets. An interesting point that someone made to me the other day is that you can think of the Société Generale deal as following the arranger model too. Basically SocGen itself provides both the collateral and the structure or “trust anchor” (basically being a heavy duty bank means they can “be their own trustee” as an Inherently Trusted Counterparty), and then SocGen Forge is the arranger, i.e. the actual point of interaction that builds out the infrastructure necessary for governance to interact with the collateral and structure.

As a practical example of the arranger model, imagine if we decide we really like working with SocGen, then in the future we may go to SG Forge and basically “place an order” for x million USD of collateral with xyz eligibility criteria, by passing a large pre approved debt ceiling to a structure where the eligibility criteria is legally implemented and SG Forge has access. Then when everything is in place from our side, and all they need is to go through legal machinery on their end, SG Forge can approach the rest of the bank knowing they have e.g. 300 million in demand from us with no governance uncertainty and no other factors that are unfamiliar to them, and that should really get the gears moving in the rest of the bank to bring us something very quickly.

“Bond guys vs loan guys”

Very happy that this perspective was brought up as it’s something I’ve also heard explained from many others. Another way I’ve heard it phrased is “does maker want to be like a pension fund or like a credit fund?”. It basically boils down to whether Maker should save money on staffing and rely more on highly standardized, liquid and “self-regulated” bonds, or have a more intensive and hands-on process where credit quality is directly analyzed and Maker can earn higher interest rates, at the cost of more bureaucracy.

I think most people inherently lean towards bonds, arguably its the simple and obvious choice - For the most part I agree, but I think there are some unique opportunities for us to go beyond the way things have been done before and this ties into the overall RWA Strategy that I believe would generate the most value for us.

RWA Strategy

One assumption I want to discuss before I get into my argument for an overarching RWA strategy is the relationship between the supply and demand for Dai. The two things aren’t independent, but rather very linked together. In the Sai days, the pitch was that it was money backed entirely by ETH, and the knowledge of and ability to verify the collateral was a key driver of demand, awareness and viral marketing for the project. On the flipside we’ve seen how much using USDC as collateral has impacted the Dai brand - although this has clearly been distorted specifically by would-be competitors that fixate on this point in an attempt to cannibalize Makers userbase (amusingly painting themselves into a corner where they’ll never be able to access RWA lest they completely sacrifice their brand). Either way, the point is simple and clear: The collateral that backs Dai has enormous impact on the brand and marketability of Dai.

So this is where Clean Money comes in - I’m arguing that the climate catastrophe and the increased awareness, panic and destruction it is causing is going to be the single most important factor that will dominate the lives of everyone for the next many decades. And with this context I would argue that Maker can utilize RWA as a tool that is less about generating a return for MKR holders, but instead is primarily focused on driving Demand, liquidity and lower cost of capital of Dai, as well as hedge the risk of a climate collapse.

My guess is that the brand effect of ETH, USDC etc is nothing compared to what actual physical objects can do, because they are much more familiar to regular people compared to abstract concepts like tokens. Or put more simply, I think wind turbines, solar farms, battery farms, nuclear power plants, agroforestry operations etc will provide us with marketing materials orders of magnitude more powerful than anything we’ve done so far, and if we combine this with a scaled up marketing budget funding a parallel, self-improving marketing strategy we can achieve massive amounts of value creation by growing the Maker Protocol and reducing our cost of capital through Dai adoption.

There is of course also the other major benefit of the Clean Money strategy which is hedging collapse with resilient assets. I’m postulating that the exposure to the marketing benefits of climate alignment is hedged by climate resilience - either the world wants to save itself and we benefit by tapping into that and being a part of the solution, or the world doesn’t want to save itself and real collapse occurs, where Maker would then be well positioned to withstand the shock and continue to provide stability.

Icebreaking the global market for green bonds and structured products

So what this means is that we shouldn’t adopt the approach of the traditional financial system as it works today or buy the same assets and do the same kind of trades that wall street is already doing. We need to use current best practice as a starting point to either help the system readjust to a new, sustainable equilibrium, or to hedge against the collapse that will occur if it doesn’t.

All in all this makes me believe that we should take an approach somewhere in the middle of the “bond guys vs loan guys” spectrum. We should focus on assets that have characteristics of bonds and structured products, but we should scrutinize them deeply, and design them more in depth than what e.g. a pension fund on autopilot would do. These assets already exist today, and are typically called sustainability-linked bonds and are basically the fixed income equivalent of ESG stocks. But it’s an extremely immature asset class compared even ESG stocks.

I believe there is an enormous potential in taking this approach, because it means we could revolutionize and “icebreak” this market by improving its standards and injecting massive amounts of supply, demand and liquidity into it. Keep in mind that while the market is basically ignored today, in the future it will be the ENTIRE market. Being the project that did the icebreaking for such a massive, historical shift, will create immense amounts of brand value.

First of all we can research current frameworks for judging sustainability-linked bonds and see if there is one or more of the frameworks that deliver real “clean money standard” meaning that it is more than just bells and whistles and has a truly net positive impact, even considering things like stuff hidden in the supply chains etc.

Then we can support these specific frameworks, promoting them with our marketing and instructing arrangers to begin sourcing large amounts of deals that fulfill these standards (initially using the “loan guys” approach), and then package them into green bonds that we use as collateral in Maker, and also promote for others to invest in.

Once we have proven it can scale, we can begin to source similar, standardized green bonds from institutions in even larger quantities. This is the moment where Maker would begin to make a real and very signicant difference regarding the climate and the fate of the planet. Once these markets, liquidity and the knowledge to trade them start to grow inside the institutions, a successful sustainability transition of the global economy becomes much more likely because the means to do it will already be there, and it will just be a matter of redirecting the cash flows away from places such as fossil fuel subsidies.

Leapfrogging black box finance by making clean finance blockchain-native

There’s one more way we can take this to the next level, and create a leapfrogging similar to when Africa skipped landlines and went straight to wireless, and that by making sure that from the start this new market for green bonds is blockchain-native.

A relatively easy first step is to simply require the arrangers not only create bonds with the collateral they create “loan guys” style, but that they also allow these bonds to be represented as tokens. This would not be a huge step since creating a bond or creating a token isn’t really that hard, what’s actually hard is getting liquidity for these assets.

The second step and much bigger challenge is to figure out the age old problem of regulated secondary markets for security tokens and other legal tokens that aren’t burdened with restrictions and bureaucratic red tape - but once this challenge is overcome, then Maker would be able to unleash a DeFi superpower to brute force liquidity into the green bond market: using green bond AMM LP tokens as collateral like the G-UNI collateral type helps provide Dai liquidity on uniswap.

If we can achieve this step, we have a real chance at kickstarting a global market because liquidity is everything, and showcasing at scale and with regulated products how much more powerful blockchain, tokenization and defi is at creating it will attract the traditional system towards transitioning to blockchain native finance (which will also be clean finance if we play our cards right).

Wall of text almost done

This is all a very ambitious vision that would take many years to carry out, but I think the most important thing to keep in mind is that all of this has to happen, and if we doesn’t we get a severe global collapse. And this ultimately means that we can go back to the climate resilient collateral as a way to hedge the success of our sustainability plans. Either the global economy transitions to sustainability and we’ll be at the forefront of it, or we’ll be the only currency to remain somewhat stable as cascading collapse occurs while we have large amounts of assets held in island nations.

Also Seb asked who the main people I have received input from in producing the OP. It’s mainly the potential arrangers such as Matt, Christian and Allan, and ultimately the model is based on the approach of the 6s structure.

12 Likes

Fully agreed with the hybrid bond vs loan approach as this allows for both scaling (institutionally sourced RWA and bonds) and learning (from analysing the specific deals). As to green bond as a token - Absolutely. These bonds can also be listed on an exchange, starting with a centralised exchange like Luxembourg Green Exchange (listing 50% of all green bonds globally) or, when regulatory issues are dealt with, on a decentralised exchange.

I would also argue that the ‘blockchain-nativeness’ should extend to underlying collaterals. Green assets financed should be tokenised as the first step and, gradually, their green impact tracked on a blockchain by IoT devices. Only when ‘green’ tracking is embedded at the asset/product level, coupled with right incentives, that we can tackle green washing. Some players are moving in this direction.. Maker has a unique advantage to pioneer this (think of smart contracts automatically reduce DAI interest rate on sustainability-linked loans once certain green impact is achieved).

1 Like

I think most people inherently lean towards bonds, arguably its the simple and obvious choice - For the most part I agree, but I think there are some unique opportunities for us to go beyond the way things have been done before and this ties into the overall RWA Strategy that I believe would generate the most value for us.

Either way, the point is simple and clear: The collateral that backs Dai has enormous impact on the brand and marketability of Dai.

At scale, it’s hard to imagine that either the bond model or the loan model would be 100% scot-free forever. Even with the best due diligence, unpredictable events can happen in the market that can threaten assets that look impregnable. With sufficient diversification, this risk is somewhat mitigated, but something that would be cool is to somehow provide some layer of equity/first loss protection to the collateral.

This somehow ties back into the governance/who vets the pool/what goes in the pool problem. Having the equity risk offloaded “somewhere else” (for example, requiring the Arranger to bear the equity/first loss risk) would ensure that interests are aligned.

2 Likes

Based on my experience every structure has its own advantages and disadvantages, but ultimately good credit underwriting is good credit underwriting.

The mistake is when we believe that structuring can overcome quality flaws, as in the RMBS happy days of the financial crisis. That is also why, in my opinion, intermediaries should be at least credit neutral but preferably credit enhancing - because their very existence is making the lender a bit further away from the ultimate borrower, reducing visibility.

Arrangers, as envisaged in @rune 's model (although we are trying to fine tune this definition) might not have the equity capital to be “credit relevant” - i.e. that equity buffer they are ready to put could be make-or-break for them as individuals, but irrelevant vis-a-vis the quantum of the loan. We should therefore ensure that those arrangers do not compromise transparency and ultimately collateral control, and be as hard as needed in DD’ing the underlying credit.

We have the luxury of not having to go very far away in the risk spectrum to satisfy our risk appetite, in my view, and that’s a great luxury to have.

4 Likes