MIP6 Collateral Onboarding Application: Stable+; US Treasury Securities Up To 12 Months Duration

1. Who is the interested party for this collateral application?

@GFXlabs; @PaperImperium

2. Provide a brief high-level overview of the project, with a focus on the applying collateral token.

MakerDAO holds more than $3,830,000,000 in minimally-diversified, zero-yielding stablecoins (USDC, USDP, soon GUSD) in its Peg Stability Modules. This presents reputational risk to Maker as a “wrapper for USDC” and existential credit and counterparty risk with regard to the issuers of Circle and Paxos (and possibly Gemini in the future). The stablecoins issued by these entities are themselves generally backed by short-dated-maturity direct US Treasury obligations, but pass on zero yield to Maker and are concentrated enough that there is existential risk should Paxos or Circle blacklist Maker’s addresses for any reason. From the perspective of capital preservation, the large exposure to USDC and USDP should be managed to lower risk of capital loss.

Unfortunately, the number of fiat-backed stablecoin issuers available to diversify this risk and still function as a stabilizer for the DAI<>USD fixed exchange rate is small. The solution is to diversify these stablecoin holdings – which are ultimately rehypothecated tbills (short-dated US Treasury debt that ranges from 0 to 12 months in duration) and cash equivalents – so that excess funds do not increase counterparty and credit risk when unneeded for peg stabilization.

Risk USDC Stable+
Can Blacklist? Yes No
Competitor? Yes No
Solvency Exposed To Crypto Market Risk? Yes No
Important Counterparty Exposed To Crypto Market Risk? Yes No
Positive Yield? No Yes
Formal Agreement With MakerDAO? No Yes
Treasuries As Primary Underlying Asset? Yes Yes

To lower the risk exposure of MakerDAO to centralized stablecoin issuers which are also competitors to MakerDAO, GFX Labs will create a bankruptcy-remote BorrowCo (proposed name “Stable+”). Stable+ will use its line of credit from MakerDAO to purchase direct US Treasury obligations up to 12 months in maturity – commonly referred to as tbills. Purchases will be made on a regular schedule with the target of having ample liquidity to replenish the MakerDAO PSMs should they experience outflows beyond a level determined by Maker governance.

This will reduce the exposure to Circle, Paxos, and (soon) Gemini by adding a fourth intermediary that holds the same or similar pool of underlying collateral, and unlike stablecoin issuers, remit some of that yield back to MakerDAO in the form of stability fees. The goal is similar to the use of the Gelato UNI USDC/DAI tokens – diversify direct counterparty risk while creating non-zero yield on what is ultimately the same underlying collateral.

Additionally, the flow of funds will be structured such that any effect upon the market price of DAI and USDC/USDP should be sterilized, preventing distortions that could simply refill the PSM, unlike direct deployment of USDC/USDP reserves in on-chain markets, which at any meaningful scale would introduce pressure on the USDC/USDP side of the peg with DAI.

The simple nature of operations should allow for low-bandwidth monitoring and evaluation by the DAO. No deal pipeline or personal relationships or confidential information will need to be managed. Data on Stable+ holdings can even be made available regularly to @Makerburn so that the public can see the exact portfolio backing the DAI MakerDAO provides.

3. Provide a brief history of the project.

This is a new project, and will only be launched upon approval by MakerDAO governance.

4. Link the whitepaper, documentation portals, and source code for the system(s) that interact with the proposed collateral, and all relevant Ethereum addresses. If the system is complex, schematic(s) are especially appreciated.

Very little on-chain activity will occur except the flow of funds into and out of the vault, which is detailed in the RWA supplemental section below.

5. Link any available audits of the project. Both procedural and smart contract focused audits.

Regular verification of Stable+’s holdings can be made whenever the RWF or appropriate MakerDAO core unit requests. All securities will be custodied directly at the US Treasury and will not involve a third party broker, ensuring the shortest possible chain of credit while still housing the collateral in an entity that is unlikely to be subject to any legal risk related to MakerDAO’s core business activities.

6. Link to any active communities relating to your project.


7. How is the applying collateral type currently used?

Short-dated US Treasury obligations are typically tbills with maturities of several days to 12 months. They are sold at a discount, and upon maturity can be redeemed for face value. They are heavily utilized as the collateral of choice in traditional finance, as they are considered to be close to risk free.

They are also utilized by corporate treasuries to hold cash, because bank deposit insurance does not cover amounts that are material to large businesses – meaning that tbills serve as a substitute for bank accounts because they do not carry the counterparty risk of a bank’s failure or decision to sever relations. Thus, corporate entities utilize tbills in order to reduce existential exposure to banking counterparties, just as MakerDAO would utilize them to reduce existential exposure to the issuers of USDC and USDP.

These two factors combine to create the level of demand that has held borrowing rates low for the US government in recent years.

8. Does one organization bear legal responsibility for the collateral? What jurisdiction does that organization reside in?

The United States Treasury bears legal responsibility for all Treasury obligations, and are backed by the full faith and credit of the United States government. Stable+ will hold a portfolio of these obligations (typically to maturity), which will be custodied directly at the US Treasury with no intermediary to minimize risk. Both US Treasury and Stable+ will be subject to US laws and regulations; US Treasury debt instruments are also issued under and subject to US laws and regulations.

9. Where does exchange for the asset occur?

Stable+ will directly participate in auctions for short-dated Treasury securities. These securities will be custodied directly at the US Treasury on Stable+’s behalf.

(Optional) Has your project obtained any legal opinions or memoranda regarding the regulatory standing of the token or an explanation of the same from the perspective of any jurisdiction? If so, those materials should be provided for community review.

Any token will simply be a placeholder to interact with the MakerDAO smart contracts.

10. (Optional) Describe whether there are any regulatory registrations for the token and provide related documentation (including an explanation of any past or existing interactions with any regulatory authorities, regardless of jurisdiction), if applicable.

Any token will simply be a placeholder to interact with the MakerDAO smart contracts.

11. (Optional) List any possible oracle data sources for the proposed Collateral type.

Aside from a variety of private and public market feeds, the US Treasury regularly updates the daily market yields on most of its securities based on auction and secondary market data.

12. (Optional) List any parties interested in taking part in liquidations for the proposed Collateral type.

US Treasury securities are the most liquid securities in the world, and ready buyers are available on the secondary market. However, liquidation could present losses if rates have risen since purchase, resulting in a lower price-to-face value of the security. Alternatively, the portfolio will target regular maturities, with a maximum time-to-maturity of 12 months on any given security, allowing for a complete unwinding of the portfolio within 12 months’ time.

The following questions apply to RWA collateral onboarding applications only:

14. Provide (a) proposed legal structure for transaction, including: type of legal entities, (offshore/onshore, form (trust, corporate, other)) and jurisdiction(s) of legal entities, and (b) likely funds flow (DAI => Fiat => DAI).

(a) GFX Labs will create a bankruptcy-remote BorrowCo, currently proposed to be named Stable+. This is envisioned to be a US-based LLC or functional equivalent, which can then interact legally to borrow from and assign creditor rights to MakerDAO’s RWA Foundation in the Caymans or some other entity of MakerDAO’s choosing (perhaps a new trust that has not yet been established). Another option is to utilize an instruction agreement, similar to what may be utilized with other traditional financial counterparties MakerDAO is evaluating.

Due to the simple, mechanical nature of operations, originations, and flow of funds, minimization of complexity risk is highly desired, but the specific mechanisms are negotiable.

GFX Labs is open to input from the MakerDAO RWF core unit(s) on alternative methods of legally structuring the agreement, but has a strong preference to minimize risk arising from complexity.

(b) Flow of funds are anticipated to be:

  1. Stable+ generates DAI from Maker vault (on chain)
  2. Stable+ swaps DAI for USDC or USDP via the Maker PSM (on chain)
  3. USDC or USDP will be directly redeemed from the issuer for USD deposited into Stable+’s bank account (on chain/off chain)
  4. USD from Stable+ is then used to directly participate in US Treasury auctions (off chain)
  5. US Treasury securities owned by Stable+ as collateral are custodied at US Treasury until maturity (off chain)
  6. USD remitted from US Treasury to Stable+’s bank account (off chain)
  7. Stable+ either directly purchases DAI on the market OR mints USDC/USDP through the issuer and swaps for DAI in the Maker PSM (off chain/on chain)
  8. DAI is used by Stable+ to repay the Maker vault (on chain)

This flow of funds is meant to minimize risk of loss in transit and only interacts with smart contracts directly under MakerDAO’s control. This flow of funds also ensures that neither DAI nor USDC/USDP are sold into the market, potentially causing distortions that could affect flows into/out of the PSMs. It also eliminates potential loss from slippage in the exchange rate.

15. Provide details on the organizational structure of the interested party, beneficial ownership, governance/control, key personnel, capital/funding resources and past financial performance.

Stable+ will be owned by GFX Labs, but structured to be bankruptcy remote from the parent company. It is likely that Stable+ will be formed as an LLC or similar structure in a favorable US jurisdiction like Delaware.

Key contacts are likely to include @PaperImperium and Getty Hill. GFX Labs will provide capital for Stable+ to invest alongside MakerDAO to meet overcollateralization requirements, and Stable+ will be responsible for its own expenses, taxes, legal, and other obligations.

16. Provide detailed summary of the proposed economic terms of the transaction, including, without limitation: commitment term, principal amount, interest rate, frequency of principal and interest payments, disbursement schedule, equity amount, funding ratios (equity/debt pro rata, equity first, etc.), collateral security, coverage ratios, currency (if not DAI) and other material terms. The quality of the proposed economic terms will be a consideration for the prioritization process.

GFX Labs is requesting that Stable+ be granted the following terms:
Stability Fee: 0.05% (fixed)
Collateralization Ratio: 101%
Disbursement Schedule: Escalating; variable dependent upon MakerDAO’s diversification needs (table provided as example only)

Depending upon Maker’s needs to diversify away from Circle, Paxos, and other treasury-backed stablecoins, the example schedule above can be accelerated or decelerated to limit existential exposure to counterparties with no legally binding agreement with MakerDAO.

The collateralization ratio is identical to that accepted by the US Federal Reserve for Treasury securities dated up to 12 months.

The current Fed funds rate is hovering at 0.08% and SOFR (the short term rates for treasury-secured debt) ranges from -0.01% in the lowest 1% by volume up to 0.05% in the 75th percentile by volume.

The stability fee is designed to provide a small net interest margin to cover the expenses of Stable+’s operations and modest return for GFX Labs to provide the service. The most recent yields on a variety of Treasury securities can be seen in the table below. Stable+ will target a range of maturities between 4 weeks and 12 months, with the primary goal being to preserve capital and liquidity for repayment should the Maker PSMs experience outflows.

Should the interest rate environment allow an increase in stability fee, that increase would only be on new – and not outstanding – DAI debt. This may require the creation of a new vault with its own debt ceiling while allowing the previous vault to be closed as older securities mature.

17. Identify in reasonable detail the risks associated with this collateral application and the underlying asset(s) and proposed mitigants (if any). The risk summary should address, without limitation and to the extent relevant: market risks, commercial risks (e.g., diversification, credit, etc.), interest rate risks, legal and regulatory risks, general industry risks, competition, etc.

Three parties would bear potential risks, and will be addressed in turn.

GFX Labs bears the risk of reputational damage in the event of unforeseen circumstances leading to some kind of losses stemming from operational mismanagement (such as DAI lost in transit or inability to purchase securities).

At 101% collateralization ratio, GFX Labs would bear capital losses before Maker. This will be mitigated by utilizing the “crawl, walk, run” philosophy as outlined by the 6S operational philosophy, and is the primary reason to gradually ramp up monthly allocations. Tbills also are considered to have zero default risk.

Stable+ bears primarily smart contract risk. There is always a non-zero chance funds can be lost in transit or due to smart contract risk. GFX Labs’ policies on fund movements on-chain follow best practices would be duplicated, as would be expected by former traders experienced with on-chain transactions. The only smart contracts that will be interacted with are the Maker vault and the Maker PSM, both of which are under the direct control of Maker governance.

Stable+ also faces the small possibility that it could be targeted as part of an enforcement action against MakerDAO, but being a completely separate entity (borrower and not core unit) should mitigate most residual regulatory risk stemming from MakerDAO.

MakerDAO’s main risk is capital loss. Should the US government or Stable+ be unable to repay their debts, Maker would experience a capital loss. Tbills are considered to have zero default risk.

Maker also faces peg risk when large amounts of DAI and USDC/USDP enter the crypto economy, but this will be mitigated by preventing any DAI or USDC/USDP from entering the DeFi ecosystem.

18. Outline the applicant’s underwriting guidelines/policies, origination strategy (marketing, sales, channels), servicing strategy (charge-offs, collections), and historical asset performance.

The origination strategy is to target a portfolio of various maturity tbills, with a maximum date to maturity of 12 months. The main risk is inflation risk. Given Maker’s near-zero cost of capital, it can lend profitably against ultra-safe collateral for any nominally positive yield. Tbills are considered to have zero default risk, though in 1979, one issuance of tbills was unexpectedly defaulted upon due to unprecedented failure of word processing programs that were necessary to compile information for checks to be mailed to individual investors. US Treasury made the payments over the following days. This is the only known default of US federal debt in more than 200 years.

19. Outline the applicant’s risk monitoring and operations guidelines/policies (e.g., charge-offs, collection, recovery provisions, data collection and technology, etc).

Only short-dated Treasury securities will be purchased with this borrowed funding from MakerDAO. The entirety of the portfolio will also be custodied directly with the US Treasury Department, making it easy to monitor and provide proof of collateral to the RWF core unit(s) on whatever schedule is required.

Because the portfolio is intended to be laddered for liquidity and held to maturity for rollover or redemption, Stable+ will not elect to utilize mark-to-market accounting, and will report all Treasury securities at face value.

20. Describe the regulatory regime applicable to the underlying asset (if any) and the applicant’s legal and compliance program relating thereto.

US Treasury securities are lightly regulated – generally requiring only KYC and cash to purchase. US government securities enjoy a number of exceptions to securities laws as well (e.g. does not require a broker to purchase). Aside from federal taxes, there is minimal legal or compliance work expected to be required of Stable+.

21. Identify any 3rd party persons likely to be relied upon by applicant to implement the transaction (legal, accounting, servicers, trustees, etc.).

Stable+ will have a limited scope of business activities and operations. This mainly consists of movement of funds, participation in US Treasury auctions, and standard business requirements such as filing taxes.

Circle, Paxos, or a future stablecoin issuer that offers redemption to US dollars will be relied upon to transition between on-chain and off-chain without risk of slippage.

Supplemental Consideration

The general parameters of collateralization at 101% and Stability Fee at 0.05% have similar precedents within MakerDAO’s system, and this proposal for GFX Labs to form a Stable+ is in line with or superior to historically accepted risk parameters.

Comparable collateral assets include the Gelato UNI USDC/DAI (0.05% pool) tokens, which enjoys a 102% collateralization and includes intermediaries of Gelato, Uniswap, and Circle. Gelato tokens currently have a fee of 0.1%, and are meant to serve a similar purpose as Stable+’s proposal – to distribute outward some of the blacklisting and other counterparty risk while generating non-zero yield. Stable+ also provides additional diversification, but rather than layering more intermediary smart contracts on top of USDC, it does so by replacing Circle/Paxos as the intermediary between MakerDAO and short-term Treasury securities.

USDC and USDP themselves, within the PSM, are treated as worth face value as well, with an effective collateralization ratio of 100%. Circle and Paxos notably hold a portfolio whose safest components are the same underlying securities Stable+ seeks to purchase. Stable+’s proposal replaces a block of 100% collateralized, 0% yield lending against Treasuries with 101%, 0.05% yield against Treasuries while also diversifying counterparty risk.

MakerDAO also lends at 0% fees against two UNI LP tokens – DAI/USDC and WBTC/DAI. The former is similar to the Gelato UNI USDC/DAI and Stable+ proposal in that it serves to diversify direct counterparty risk. The latter boasts a 140% collateralization ratio but still provides no yield to MakerDAO to compensate for any risk or opportunity cost.

Like other real-world assets approved by MakerDAO, Stable+ will utilize a “mint first” approach that relies upon legal safeguards to draw DAI to purchase the underlying collateral and/or develop it to unlock value. In this case, it is less complex than existing real-world asset arrangements. Stable+ relies upon the inexhaustible supply of US government debt securities and the short credit chain of MakerDAO>Stable+>United States Treasury to ensure drawn DAI can be deployed in a timely fashion.


Interesting proposal–simple questions (Kind-of reviewed quickly and might have missed some details).

One detail that is not hard to miss is the requested SF. The Industry average expense ratio for well established US Treasury funds is currently around 0.60%–being that you folks are new to this–don’t you think the SF be somewhere close to 100 basis pts? Short-term rates seem to trending higher in 2022, as well as demand for long-tail (30-year is now over 2.1% and the 10 is hovering under 2%)

Also, will Stable+ be supervised under a DST?


Interesting concept.

some questions:

(1) Can you provide a diagram with entities and flow of funds?

(2) How do you intend to manage duration risk for the longer stuff? What is your interest rate management strategy?

(3) How will you manage Maker’s liquidity needs? In stressed times?

(4) What are the backgrounds of the folks who plan to manage this treasury portfolio? Particularly what is their experience in managing high quality bond portfolios? Do they have a track record?

(5) How should Maker compare this with say a plain vanilla ETF option such as Vanguard Short-Term Treasury ETF (VGSH)? This fund charges .04% fees


Great questions!

A fund has significant regulatory overhead and filing requirements. Stable+ would not securitize the underlying or otherwise do business with the public. Also, transactional costs are anticipated to be held down by 1) holding to maturity, 2) custody the securities directly at US Treasury, which is free, 3) participate directly in auctions, which bears no transaction fees.

The 12-month tbill currently yields 41 bps. Even if all of the portfolio from this proposal was in 12 month tbills, that’s simply unworkable. Remember also that there is a margin of overcollateralization so even if Stable+ had zero costs whatsoever, the 41 bps could not be passed on to Maker. There is a sister proposal to target the middle of the yield curve that provides significantly more yield back to Maker than this one.

If that is helpful, we can provide one. The flow of funds is very linear, however, and only involves Stable+, Circle/Paxos, and US Treasury. Please remind @PaperImperium if you still want a diagram.

Duration risk is envisioned to be managed by a fixed rate offered by MakerDAO. If rates need to rise, the intention is to have a separately accounted for vault/line of credit at the new rate. That way bills can be held to maturity (maximum of 12 months) without MakerDAO or Stable+ bearing duration/interest rate risk. Because MakerDAO has nearly zero cost of capital and the bills act as a replacement for excess USDC/USDP, which yields and will always yield 0%, it is a strictly better return, even in a rising rate environment.

Because the proposed schedule of increasing DC is not set, Maker has a month-to-month flexibility in how much to add to this portfolio. Ultimately, tbills are a “parking place” and if PSM inflows reverse, pausing new allocations is appropriate. Due to the short duration, the entire portfolio can be repaid in a manner that replaces USDC/USDP over a maximum of 12 months. Maturities will be a range, and laddered, with substantial liquidity (minimum target 1/12th of the portfolio) available to make repayments. Maker will need to communicate its needs clearly when it desires unwinding of a portion of the portfolio due to insufficient funds in the PSM. The example schedule of allocations included in the proposal would only amount to 10% of all USDC in the PSM currently.

GFX Labs is founded by traders who are familiar with best practices in fund movements and market operations. @PaperImperium has participated in US Treasury auctions. The proposal is designed to minimize complexity – buy-and-hold short term Treasury bills as a way to diversify credit risk away from Circle, Paxos, and any future stablecoin issuers, which utilize the same underlying collateral. There is no intention to actively trade the bills, and they will reside at US Treasury and perpetually rolled over until redemption is required. No additional holdings outside of direct US Treasury obligations are planned.

Maker should probably explore multiple methods of diversification, and multiple proposals, given the sheer size of exposure to USDC and USDP relative to Maker’s current DAI reserves. Maker would need to create a controlled legal entity to purchase that or any other ETF, however.


My understanding here regarding structure of RWA is:

  1. collateral comes first
  2. then DAI is minted.

I think we can fund this beast initially out of the SB to get it rolling. I think I don’t like the idea of waving the above process because it will probably set a bad precedent.

I need to review this more carefully but the above first popped out at me.

1 Like


Maker will likely need a good deal of liquidity if another Crypto Winter/tail event arrives. Can these funds be accessed quickly (say over a few weeks) if things get ugly? Or should be view them as less liquid than the current stable coins (which I dont love)?

Also could you please post bios of these folks? with details on the types of bond portfolios they have managed historically?

“GFX Labs is founded by traders who are familiar with best practices in fund movements and market operations…”

1 Like

Not a big fan of this proposal, especially in its current form. Let me preface this by saying that I know Paper is doing this because he thinks it is helpful to the DAO, and he has been a longstanding and passionate member of the community. The issues that I’m pointing out, although I believe they are critical and need to be treated as “red lines” that should never be compromised, have not be formally adopted by the DAO - but that’s not a good reason for ignoring them, instead it’s an argument for moving forward with adopting enforceable minimum standards that safeguard Maker’s interests.

  1. This should be treated as an arranger, and be required to follow best practice as such. At minimum it should require a top tier, independent trustee that is capitalized, regulated and reputable, to safeguard Maker’s interests.

  2. There’s obvious conflict of interest and separation of powers concerns having a delegate also be an arranger. Doesn’t mean GFX labs shouldn’t be allowed to do it, but they should choose whether they want to run a recognized delegate or an arranger.

  3. becoming an arranger shouldn’t just be a popularity contest, it should never be rushed but needs to be a long term process, and new arrangers should be at the same level of existing arrangers in terms of professional experience, relationships etc.

MKR holders should always demand the highest level of quality and security when it comes to sending large amounts of money to centralized counterparties. There is simply no reason to compromise on this.

Paper already approached me with this idea and i explained the above concerns. But I guess he then chose to just go ahead with it and not even address them, which is a red flag IMO. At least this is not a MIP, just a MIP6 application, so it’s not something we have to urgently react to and we can use this as an opportunity to discuss how to make the overall RWA process safe without requiring MKR holders to always be on guard.

IMO this serves as another example that we need to push for enforceable standards for critical principles like separation of powers and best-in-class recourse from RWA counterparties. They are convenient to ignore, but once we get in the habit of ignoring them it will inevitably build up massive hidden risk.

We need guarantees that things that contribute to such risk buildup cannot even go into the process or be pushed by the facilitators. MKR holders shouldn’t have to always be on guard for such proposals.


Interesting proposal.

  1. The legal safeguards in the event Stable+ misbehaves are unclear.

  2. Having Sable+ be a non profit that only pays operating expenses (transparently) and regularly returns the excess gained from the yield-SF spread would be much closer to the spirit of a simple pass-through entity.

2b) What is the advantage of doing Stable+ compared to having eg Growth task an established trusted entity to follow the simple dai<->bills procedure after going through some “using the blockchain” training?

  1. The team background needs to be much more spelled out. For instance the sentence “Paper has participated in treasury options” would have felt more honest if followed by “recently, for the first time” (from public discord communication), and I assume for the purposes of writing this post. Could you give much more background?
1 Like

This is a very interesting proposal. It brings many questions to mind, given my background in Treasury Management (12 years at the Bank of Canada, buying t-bills and other highly-rated debt for Canada’s foreign reserves portfolio).

What is the proposed mechanism for these purchases? Sounds to me like it plans to use TreasuryDirect, the Treasury’s retail-facing portal. If so, why wouldn’t Maker use it under their existing legal entity? If not, it would require a back-office team, middle office reporting, and legal/regulatory reporting functions, as well as front office traders. Or someone to handle those functions on your behalf. Is GFX Labs proposing to offer these services? TreasuryDirect only allows non-competitive bids, though that’s sufficient for a passive strategy, it disallows more interesting multi-bid strategies, which would also fit a passive mandate. I think setting up a detailed investment mandate would be necessary to specify such details. The proposal says the portfolio will “typically” hold these to maturity, clarified with “no intention to actively trade”. Under what scenarios would that not hold? TreasuryDirect does not allow you to sell your holdings in the secondary market. How would this be accomplished? It would require direct relationships with brokers. What type of maturity laddering will be used? Strict laddering implies only purchasing year bills, but those mature only once a month, meaning you would not have access to funds for up to a month (without secondary market selling).

This brings to a few risk factors that aren’t really addressed. Mark-to-market risk would be introduced from the price volatility of t-bills. Yes they’re one of the most boring rates in the world, recently locked around 0.05%, but risk of loss is possible. If liquidation would be required shortly after the Fed raises rates, selling bills purchased at 0.05% but sold at 0.25% increments higher per Fed hike would incur losses of 0.125% per Fed hike, assuming an average portfolio duration of 6 months. The proposal to ignore mark-to-market risk is in-line with the current practice of fixed NAVs for government Money Market Funds, but that’s largely a historical artifact that doesn’t reflect best practice to value tradable assets at market value, and would make it impossible for Maker to monitor for such loss. This mark-to-market risk would originate from rates rising, so duration and interest rate risk management would be key to avoid this. Valuing a portfolio of zero-coupon bills, which trade at a discount to their face value, at their face value means systemically overvaluing this portfolio, resulting in systemically missing the proposed 101% collateralization target. Although this target could be increased, a 1% discount would make the current position undercollateralized at the current level, which would occur if rates rose to 2% on a portfolio of 6-month average term. Additional liquidity risk would exist in times of market stress, as recently experienced in March 2020 as markets priced in a global pandemic, requiring Fed intervention to fully unseize markets. Is that the type of risk Maker wants to be exposed to?

Settlement and liquidity risk is introduced by the additional delays incurred from the settlement periods involved in cash transfers and investment settlement. Liquidity risk is the risk of not having liquid funds available when required. How long would it take from acquiring funds from Maker to fully investing them? Would this be on-demand, or only on a predetermined schedule, say to match year bill auctions? The auctions have a 2-3 business day settlement lag, with a 4- and 8-week bills crossing a weekend. Sales typically have a 1 business day (t+1) settlement lag. They can be sold for same-day settlement, but with strict intra-day deadlines to meet. These are the types of operational inefficiencies that crypto is well set up to solve. I don’t like the concept of returning to them. An ETF like Eumenes mentions avoids having to deal with these details yourself, outsourcing them to an investment manager whose ability to spread fixed overhead costs to a much larger asset base I believe outweighs their higher regulatory and filing requirements. They also have an established track record, detailed regulatory oversight specific to money-market funds, and access to a wider set of investment options, like repurchase agreements or the Fed’s reverse repurchase operation which acts as a floor to money market rates but is available only to some of the largest funds (list) which is one reason why many t-bill rates trade below the floor.

Counterparty risk seems to be the main advantage presented, yet it’s unclear how much that risk is actually reduced. One could argue you’re increasing counterparty risk to this new investment entity with no track record. The entity is describe as bankruptcy-remote. Doesn’t that mean any losses they incur won’t be passed on to GFX Labs? GFX Labs would put up capital to achieve the overcollateralization, so they’re investing alongside Maker, within the Stable+ structure, while also owning Stable+. “GFX Labs would bear capital losses before Maker” sounds like GFX is offering protection against loss, whose value depends on GFX Labs’ counterparty and credit risk. It sounds like the two entities are unavoidably interlinked. The proposed flow of funds still goes through USDC or USDP, going through the centralized counterparties you’re aiming to reduce counterparty risk to, in order to redeem for face value. Could that step be improved upon, for instance by giving a DAI loan to GFX Labs for the working capital? That would further lower counterparty risk to Circle and Paxos, while making the counterparty risk to GFX Labs more evident. The diversification benefit itself is uncertain, as the investment relies on faith in the US Treasury, and increasingly the ability of deadlocked US political parties to pass debt ceiling extension bills to avoid technical default, around which liquidity in Treasury Bills suffers. The counterparty risk of GFX Labs and this new US-based entity seem subject to the same, if not more, regulatory risk than the established Circle and Paxos with more industry support and larger legal and lobbying power.

I second @rune’s point about separating the decision to enter a new asset class vs. hiring a manager for that asset class. While it’s common for an asset manager to pitch a new portfolio allocation, it behooves an investor to do due diligence and shop around for competing investment managers before hiring one. It would be cleaner for Maker to evaluate the two decisions separately, without intermingling the pros and cons of the asset class versus the proposed implementation. Would MihaiCo be able to pitch for this business? I’d favour Crypto-World Assets myself. I have to admit I haven’t followed closely the discussion of Maker’s venture into Real World Assets. Though it seems to me the crux of the argument was in finding sufficient depth of liquidity. I hope to grow the capability of DeFi to replace the role of inefficient and low-return assets like t-bills in institutions’ portfolios. Growing that liquidity in stable conservative forms is essential to that end. Maker could play a large role in supporting such budding efforts. Element Finance, which I have recently joined, currently has 64 million of liquidity in their USDC principal token pools, offering fixed rates of 3-5%, priced the same as t-bills. I would be happy to discuss further :slight_smile:


@rune , thank you for taking the time to read this application so quickly. We had hoped it would spark your interest. As the most successfully decentralized DAO in DeFi, Maker is used to setting precedents, and it’s important to execute well.

Agreed. This is in part modeled on what is being offered by Monetalis’s application, which was the source of much inspiration. Mirroring Monetalis’s example, a brand new legal entity (or even set of entities if complexity is required to interact with Maker) will be created when the first steps of approval are taken. A trustee entity of Maker’s choosing or a new one can be created if Maker desires an additional counterparty to enforce its rights as a creditor. The specifics have been purposefully left vague in order to meet the needs of Maker, and the RWF core unit has generally made clear they do not engage with applicants prior to greenlighting. The hope was not to wed the concept of the proposal to a specific set of legal particulars when no input from Maker was possible – we want the proposal to be evaluated on its economic merits, and can package those in the legal structuring that Maker governance is most comfortable with.

GFX is also offering its own capital to satisfy the overcollateralization ratio for MakerDAO. This is again modeled on Monetalis offering the owners’ equity in the new Monetalis legal vehicle as security.

You have expressed much support for that application, and GFX hopes to adapt the same philosophy for Stable+, with similar capitalization, levels of regulation, and going a step further with even more easily monitored collateral that is also considered the safest in the world.

GFX has demonstrated a willingness to roll up its sleeves and dig into MakerDAO to provide value such as MIP62: Collateral Offboarding Process, a request to extend D3M to Compound, a request to raise the threshold of the Emergency Shutdown Module. This particular MIP6 collateral application is designed to further de-risk Maker from having existentially threatening exposure to its chief competitor. Being an active participant in governance of the protocol includes building solutions for that protocol.

GFX is confident that the experience of its founders is relevant to safely managing the flow of funds on chain, and who have three years of experience as traders at Grapefruit Trading, a subsidiary of 4170 Trading. Unfortunately, most of their past work is confidential, but one of the public achievements is their early participation in WBTC and becoming a key minter.

We have also already been cleared to participate in – and have participated in – the auction process with the US Treasury, including the use of the Treasury’s services as custodian of the securities. Noncompetitive bids (which guarantee purchase) can be submitted to a maximum of $5,000,000 through TreasuryDirect, and competitive bids can be submitted to a maximum purchase of 35% of the individual offering (which is typically in the billions) through Treasury Automated Auctioning System (TAAPS). All auctions are open to the public.

TreasuryDirect as a platform charges no fees, and TAAPS fees are low. In the event that TAAPS access is not approved for Stable+, transaction fees related to tbills often are negligible-to-zero. Until such time that competitive bids from Stable+ are of size to affect yield at issuance, the cost structure of using a brokerage is materially the same as directly accessing TAAPS (though direct access to TAAPS is the first choice for the new Stable+, purely to minimize the number of intermediaries whenever possible).

GFX fully agrees and is ready to receive input on what structure would be acceptable to community members. The hope is that an agreement similar to that being considered for the SocGen proposal can be utilized, or the DAO appoint a legal representative such as a trust the DAO directly controls to serve as counterparty to an agreement with Stable+ similar to that utilized in the Monetalis proposal and the 6S structure.

This proposal was designed to limit the number of counterparties with even minimal exposure to the process. This cannot, of course, be completely removed for obvious logistical reasons – even 6S relies upon Genesis and Monetalis relies upon AWS – but in this case are limited to Stable+, Circle/Paxos, a US bank, and US Treasury. A broker would be also be included if unwinding on the secondary market was necessary or TAAPS was not accessible in time to deploy large bids. Required trustees are assumed to serve as counterparties to legal agreements, in the same manner that they are being utilized for existing real-world asset collateral financing.

The proposal is generally agnostic to the particulars, but presumably Maker would wish to utilize a trust it controls as either a legal counterparty as creditor, or as both counterparty and intermediary. If a particular structure is preferred, guidance will be need to be provided by the RWF CU. Typically there is not much interaction until after a green light vote. If input is available prior to a green light poll, that is even better and much preferred.

Given the size of exposure to USDC/USDP, Maker should explore many methods of diversifying that exposure. Anyone interested in operating in a nonprofit manner may find some useful information in a prior proposal by @PaperImperium. In order to provide capital for overcollateralization, GFX would not be in a position to form Stable+ as a nonprofit entity.

Given the size of exposure to USDC/USDP, Maker should explore many methods of diversifying that exposure.

Participation is simple and open to anyone. It does not require any special skills beyond a short KYC form for noncompetitive bidding (up to $5,000,000 per auction). Approval for TAAPS (to bid competitively) requires slightly more paperwork in that it is more than one form and requires a bit more work to arrange for custody, but is also open to the public (including to private individuals if anyone is interested). There are typically nine tbill auctions in a month.

@PaperImperium served as the general partner of Arbor Lea Partners, LP from Dec 2010 to Jan 2013, which served to manage family investments. The founders of GFX each have three years of experience in trading onchain at Grapefruit Trading, as detailed briefly above in this thread. Transfer of funds on chain is identified as the highest risk in the operation.

TreasuryDirect would likely be utilized to ensure a minimum allotment (up to $5,000,000 per auction). Beyond that, the primary choice is access through TAAPS. Should that prove impossible for Stable+, a brokerage would be used. Transactional costs are likely to be lower with a brokerage than through TAAPS – though the costs are negligible – but the preference is direct access to minimize intermediaries and ensure access to unlimited stock of securities. Unless allotments are so large as to influence the yield, there is likely no material difference to the returns, even at the scale of the example in the proposal.

That would be the easiest. However, Maker does not have a legal entity.

The “typically” applies to when Maker requests unwinding of positions. Otherwise everything will be held to maturity.

Strict laddering of 100% allocation to 12-month bills is probably not ideal, particularly for the first year. The goal would be to have access to at least 1/12th of the total liquidity at any time within 30 days, so the basket of maturities will change based upon the rate of monthly increase/decrease. If allocations are increasing in cadence, more weight will need to be placed on shorter duration bills, while a slowing of allocations may allow for more of the longer duration bills (depending upon upcoming maturities).

At large scale, it would likely be appropriate to set a fixed amount available within 7, 14, or 30 days that can be used to reverse the flow of funds quickly to replenish the PSMs. The largest 30-day outflow has been around net $800m (from mid Oct 2021 to mid Nov 2021). The scale for that liquidity strategy is unlikely to be achieved in the short term, however. We are happy to take guidance from the community if there is a desire for a specific level of liquidity from the outset.

Because Stable+ would not be engaged in business lines beyond holding US Treasury debt, and not engaged with lenders other than MakerDAO, mark-to-market does not reflect the risk exposure of Stable+ or MakerDAO. The purpose of Stable+ is not to maximize yield through trading – which introduces new risks – but to lower the risk associated with existentially large, concentrated exposure to issuers of competing stablecoins. Remember also that Maker is unique in that its cost of capital is generally decoupled from interest rates because it does not depend upon sourcing funds from a depository base. Rising rates may present an opportunity cost, but do not present a risk of capital loss unless the collateral fails.

Generating DAI and swapping through the PSM is negligible. Redemption of USDP/USDC is in theory a negligible amount of time as well, but in reality will depend upon a wire transfer to deposit into a US bank account, which should be assumed to take 48 hours to fully settle. The date of settlement for US Treasury auctions is the date of issuance, and will be 3 to 5 calendar days after the close of the auction, with weekends or holidays being responsible for issuance more than 3 days after the close of an auction.

Fund flows would generally be planned around the auction calendar.

Maker should explore ETFs and other methods of diversification, though a legal entity of some kind would need to own the shares of the ETF. Maker could lend to a borrower that holds the ETF, but would need to be comfortable with the price fluctuations of the ETF, which has been around 1.6% for the one Eumenes mentioned as a possible alternative. The primary service we are providing is not in the procurement of tbills, but rather the entire process from working with an entityless DAO, moving DAI on chain, converting to US dollars, and all of the legal/accounting work required to establish and operate Stable+. The business model and collateral value is generally not sensitive to whether the tbills are bought directly or via a traditional intermediary.

The goal would be to make GFX and Stable+ both remote from each other, such that the creditors of GFX Labs would be unlikely to reach Stable+, and vice versa.

In that failure to repay Maker would typically result in recourse to Stable+ and that capital invested in it, yes.

This would lower the exposure through the transaction, rather than continue the exposure through holding. While not the intention, a pattern of delays in redemption could also serve as an early warning to Maker to find ways to diversify its holdings on-chain quickly.

Agreed. The more the merrier, and the more participants looking to improve protocols the better. Note that this proposal itself is designed to compete with USDC/USDP which are (in theory) just Treasury bills stripped of yield.

Please do. One additional entity will not be enough to spread the counterparty risk broadly enough.

Bingo. It’s hard to find a parking place for $3.94b in stablecoins (there have been more than $100m of inflows since this proposal was posted yesterday).


Thanks for this @PaperImperium. I like the idea of reducing the risk of relying on centralized issuers while simultaneously gaining direct exposure to yield opportunities. I will also echo some of the sentiments about safeguarding the DAO because we’ll need some form of transparency in ownership and solvency since this will be directly credited by Maker.

Would it be possible to send this available data on-chain before minting DAI to alleviate active monitoring? For example, Stable+ posts on-chain a proof of assets to confirm that there’s sufficient collateral to mint DAI?