Investing our liquidity in short-term ETF, managing PSMs exposures



After polling on the idea to invest in financials assets here, it was incorporated in the RWF strategy (the Scalable off-chain investments bucket). It was also discussed with the RWA Committee this week. This post provides more information on the current status.

The idea

Even with the bull run resuming, we still hold more than 50% of stablecoin on the balance sheet. This is a strength for the peg, but doesn’t generate any revenue, is a concentration risk, and is bad PR.

Institutional vaults and “regular” RWA might, at some point, solve the problem but it is impossible to expect perfect balance sheet management with such tools. You don’t fire DAI supply in short term. We are extending the credit line so we also need a buffer.

Therefore, we will always need a buffer of liquidity (defined as any assets that can be sold in short term to keep the peg (from a few minutes to a few days).

The final idea would work like the diagram below. For each fiat-backed stablecoin PSM we would define a target DC (let’s say 3% of DAI issued) with a min/max range (2.5%-3.5% for instance). When the DC of the PSM is above the the max target, the excess (possibly converted in DAI) would be send to a cryptobroker to be converted to cash, send to a stockbroker and invested in a ETF. We would increase a ETF-A vault in our system. The ETF would be a short term investment grade bond.

The opposite can be done when the a PSM is below a min target. This might require human judgement (MKR executive).

This way, we limit our exposure to fiat-backed stablecoins to a defined level. There is a counterparty risk with the stockbroker and the ETF. Nevertheless, even $1B would be small.

Were are we?

Legal structure/KYC/AML

The idea is to leverage the work done on the RWA Foundations. The first, RWA Foundation, used for SolarX, is incorporated in Caymans and all the corporated servicers have been appointed. They are currently working on getting a KYC/AML package for US Bank (which is not used here). If we pass US Bank KYC/AML we should be good to move to other providers.

There is still the formal Articles of Associations to be implemented and defined the MakerDAO/RWA Foundation communication processes.

The final setup will have a to be rubberstamped by a global law firm.

The legal entity would have nothing to do on the DAI → ETF way (beside accounting). For the other way, I’m not sure a stock broker will be able to take the decision itself or take direction from a third party (i.e. MKR executive). So the legal entity will have something to do.


We have two leads but waiting on the KYC package.

I’m still trying to find a global servicer provider that could take DAI and deal with everything. So far no luck.


Not much work here aws being done. Waiting on the KYC package and the actual ETF selection to find the best partner. The ETF providers are able to introduce us.


I had discussion with VanEck and Vanguard at this stage. They all favor the use of an ETF (versus a managed account) for simplicity and cost.

Both are low duration risk (duration < 2 years), with exposure mainly US (but not only which adds diversity). Both are able to provide 100M of volume over a day if we go through a market-maker like Jane Street.

The proposed products are the following:

In order to add an ESG product:

So it’s about scale, reducing/managing the USDC exposure and getting some revenues (but not much in the current environment).

The risk is up to a 5% drop which was seen during the Covid crisis. It was not that bonds defaulted but credit spread increased and there was strong liquidity issues. In normal markets, not much happens. This remains to be analyzed.

We can go higher on the quality ladder but we would get nothing in return (there will be structure/trading fees).


The poll was to go forward with MIP21. I still have doubts it would work at scale, but we can start experimentation with that. I’m drafting a study on how to fix this problem (at least to frame the problem).

Next steps

Assuming there is no strong opposition, we will continue working to move the project forward.

Feel free to provide comment and ideas.


I recently listened to a podcast about custom ETF baskets. It sound like custom ETFs could be a way to filter by our ESG criteria, but I have no idea if any large players are offering custom bond ETFs at this time.


Is it fair to say that we are substituting exposure on a money-market type of security (USDC) with another money-market type of security (short term high quality bonds thru an ETF)? Still, I believe there’s a benefit in terms of extra (small) yield, some currency diversification, and reduction of counterparty risk.

It would be still interesting to read a risk/ return analysis that includes duration risk (with rates where they are even short term floating bonds run duration because of the delta effect on the principal repayment), liquidity, and complexity.

I believe that an efficient use of maker’s balance sheet / available liquidity is welcome. But aren’t there crypto-native ways of using such a liquidity? Or wouldn’t those help the peg mechanisms?

Looking forward to hear from the team, keep up the good work guys.


Investing some of the surplus into ETH itself can be a good idea too since we all believe it’s an asset that holds value very well.


Thanks for putting this together Seb. Would love to see the DAO start leveraging it’s balance sheet and putting the large amounts in the PSM to productive use, while also reducing counterparty risk. What are you thinking of in terms of initial amounts for these ETFs?

I would also love to start experimenting with DeFi native yield opportunities (depositing USDC directly into AAVE, Compound, yearn, etc). Of course these come with different risks, but we can start small and scale over time and earn significantly higher returns.


Custom baskets can be expensive. But if you’re strictly reducing breadth of investments to a specific set of ESG criteria, may be the best option to stay adherent. To Luca’s comments, sounds like there was drawdown analysis done for the ETFs selected over the past ~2 years (mainly dealing with March 2020 reaction) - suppose it matters if the liquidity is expected to be used in short-term or if it’s more expected to be accumulating unless some outside scenario determines it would be needed (thus the short-term nature).


It will be gradual starting with 1M. Currently, the upper limit is $1B assuming a max drawdown of 5% (as the Surplus Buffer is a bit above 50M).

Obviously, we can go lower on risk to have more exposure but that would lead to quite the same result at best in term of revenues. We could also avoid using mark-to-market but it wouldn’t be good for DAI transparency (even if mark-to-market is far from perfect).

1 Like

USDC is safer on the investment side, less on the operational/regulatory side.

Crypto-ways are difficult to implement. You can buy aDAI but then you are kind of delta-edged at scale so very little impact except on borrow rates.

There is no safe $-denominated asset in DeFi beyond fiat stablecoins (and their lending derivatives).


The etf return seems to be miniscule and not worth the effort. Makerdao is better off depositing the usdc into celsius, coinbase or one of yearns protocol for greater than 4% apr.


the approach seems sensible.

A few observations

  1. All three ETFs mentioned are under a billion dollars in assets. Would Maker’s potential size have a real impact on their portfolio?

  2. The Vanguard ETF has almost a 2 year duration.

  3. PIMCO has a decent short duration bond EFT (MINT). I has about $14B in assets and .5 year duration and similar credit quality with the other ETFs. Maybe worth considering?

Here is a performance comparison graph for two of the ETFs.
The last 3 years of performance by FLTR (Van Eck) and MINT (PIMCO) are below. I could not get the price feeds for the Vanguard one to work.

The March COVID impact on FLTR is about down 7% while around down 3% for MINT.

  1. All three ETFs mentioned are under a billion dollars in assets. Would Maker’s potential size have a real impact on their portfolio?

For the first 2 I spoke with the asset managers and there wouldn’t be any impact if we use a market maker on the primary market. Liquidity could sustain 100M volume in a normal day.

  1. The Vanguard ETF has almost a 2 year duration.

Yes, there is a tradeoff between duration and return on investment. The main problem is that ultra safe asset are not yielding much and I expect to have at least 10bps of structural fees + trading fees.

2 years gives us some reward and provide a buffer (relative for sure) in case yield are going negative on the short term front.

  1. PIMCO has a decent short duration bond EFT (MINT). I has about $14B in assets and .5 year duration and similar credit quality with the other ETFs. Maybe worth considering?

We sure could consider it. My concern would be that it’s an active ETF with quite a lot of fees (0.35%). The use of “enhanced” and “active” is a warning. They are comparing themselves to the FTSE 3-Month Treasury Bill Index which is really a joke. The 0.97% yield to maturity is super attractive, but I have no idea how they could have that.

I’m more leaning toward simple beta for such investment.


What a nice idea, I’m not big fan of massive amount of liquidity just relaxing and earning 0%.

But I’m not sure about the ETF approach. Why are we choosing ETFs? Then, why are we choosing bonds? 0.6% yield I don’t think gets the impact we could have with another focus.

What about high impact investments and using these cash for creating value for both, Maker and real world. Just an idea, but what about investing in small, high-growth startups, and creating a “startup accelerator base with liquidity from Maker”.

I know that it could be skyrocketing the risk and the complexity thought from the bonds ETF idea. But it’s that, just an idea that Maker can start with very little exposure and a higher return in the long term in earnings and real-world impact.

I think that these great ideas from the Maker team could be used for great impacts! :sunglasses:

1 Like

Yield is really a secondary concern. These are our “foreign currency reserves” that are needed to defend the DAI <> USD peg. We really cannot afford even small risk of losses, as they already back existing DAI. They also need to be extremely liquid so that we can regularly sweep back and forth between the PSM and the reserves (perhaps ETF, perhaps T-bills, perhaps something else).

We want a non-zero yield because we have so much operating leverage. But a lot of the motivation is to reduce credit risk concentration to USDC.

At the end of the day, it needs to be extremely liquid and extremely safe. That’s not a large universe of investment choices, and I suspect the ones @SebVentures has suggested will be our best choices, or at least very similar to our best choices.


Thanks for the explanation, now the rationale is more clear to me. It makes all the sense.

After thinking more about that, I would like to ask two things

  1. What is “ETF-A” and its goal here? Is this a new type of vault or a tokenized reference of the amount of ETF that Maker has? I’m trying to figure out ways in which this can be as transparent as possible, taking into account that money is being taken from the Maker protocol in cash to the real world as a final destination.

  2. If the concern is on the USDC/USDP exposure, should we be thinking of this as a short-term solution, while we think of a long-term, root solution (replace/fix PSM) as the idea proposed in [Signal Request] High Leverage G-UNI Vault to Replace PSM-USDC Exposure?

Sorry if I’m asking answered questions, I want to understand everything I can! Thanks again :grinning:


We don’t know yet, but I would assume it would look similar to our existing real-world asset vaults — except we could use the market price as the oracle instead of manually monitoring.

Well, we will always need some amount of “dollar proxies” to offer open market operations. Maker actually executes this much better than any major central bank. Unless the US Fed or Treasury launch a digital dollar on Ethereum, we will always need some kind of redeemable “dollar coin” to arbitrage against.

But we certainly don’t want to have half the DAI supply in zero-yielding reserves.

It also probably behooves us to figure out a way for PSMs to generate at least minimal revenue in the event we need to spin off an entity to do that at arm’s length. The AML Act of 2020 is currently without rules, but includes the exchange of digital currencies in an expansion of the Bank Secrecy Act. And we must at all cost find ways to ensure that DAI does not legally fall under a requirement for KYC. Hopefully it won’t come to that, but it’s important to think a few steps ahead.


@tobalgarcia , the risk/reward balance is hard to define but this thread is really about the second line of defense on the DAI peg (after the USDC/USDP/GUSD PSMs).

PSMs are real-time, don’t yield anything, and face an undefined counterparty risk.

Those short-term ETF would be the second line of defense for the peg. I had some talk with tokenizer startups, but really the regulatory uncertainty is a real problem, and such an ETF is not at all the priority for them. Therefore we will use the TradFi route which at least is proven for decades. It will also give us building blocks for the future.

The third line of defense is the rates on the crypto-back loans. The last line of defense is institutional vault and/or the other RWA.

I understand some think the yield is way too low, some think the risk is way too high. I’m also concerned about all the structuring costs and the complexity of the matter. I’m also thinking of doing one US route (the VanEck or whatever else) and one Ireland route (the Vanguard one) to get maximal redundancy in case it’s needed.


Love the bonds idea!

Curious if you have thoughts on how those funds would be used to recapitalize the protocol should the need arise.

For instance in another black thursday style event where you see the Stability buffer go negative. Would the RWA foundation be legally obligated to do anything with these assets? What would those obligations be?

Further what happens to these assets in the event of ES? Can’t exactly redeem dai for 1/25 a share of FLTR, so again how would these assets be managed in that type of situation?


We are extending a loan to the RWA Foundation so it’s not an additional buffer in case we need it.

The RWA Foundation will take direction from MIP21 so if we liquidate the loan that finances the ETF subsidiary, they will have to sell those to repay the loan. It’s working like any RWA asset regarding ES. They will also take direction from DAO Resolution issued by MKR Governance.

1 Like

Hmmm, a few points:

USDC is utterly centralized but has the convenience that it offers its money-market-fund interests in the form of tokens that are quite liquid on-chain.

ETF shares are utterly centralized but don’t offer such a convenience.

Is the plan to tokenize the ETF shares through the relevant RWA Foundation?

Even if the plan is to tokenize the ETF shares through the RWA foundation, tokenized centralized ETF shares are inferior to the tokenized centralized money market shares called USDC in several noteworthy ways:

(1) tokenized centralized ETF shares are not highly liquid on-chain, whereas USDC are

(2) tokenized centralized ETF shares will be centralized under the management of a completely unregulated and untested foundation team, whereas Circle at least has a lot of eyeballs on it and a long-ish operating history

(3) the blacklisting of tokens is a red herring and the real risk for both USDC and tokenized ETF shares is essentially just counterparty risk (blacklisting is a subspecies of counterparty risk)…the counterparty risk is higher, not lower, for a new RWA Foundation than it is for Circle (see point (2))

(4) although ETFs can grow in value (unlike USDC) and thus offer ‘yield’, USDC also offers yield opportunities (and have the advantage of being on-chain).

USDC dominance in MakerDAO collateral is really bad (because of its centralization/counterparty risk) but basically just trying to turn MakerDAO itself into the new Circle by standing up an entity that stabilizes the peg in the same custodial-money-market way as Circle and Tether do seems like a very weird way of addressing the downsides of USDC within MakerDAO.

1 Like

On the USDC vs ETF, I think this is quite explained in the thread. None are money-market-fund interests. USDC is an on-demand liability of Circle yielding nothing and the ETF presented are short-term investment-grade bonds.

The point about ETF being utterly centralized is kind of moot as this is the direction MakerDAO is going anyway. You do RWA, you have a link to the real world. Is Circle a better counterparty than Blackrock? Is an ETF more robust than a stablecoin? Those are the questions.

It’s not a plan to tokenize anything.

On the RWA Foundation, the counsel of NebulaCapital and Monetalis ended up with the same structure i.e. having a legal entity managed by servicers that act on MakerDAO wishes.

In the case of ETF, those servicers could be JTC or Intertrust (keeping the Foundation in good standing), Genesis Trading (DAI <-> USD), Jane Street (primary market making) and BlackRock (ETF management). All wired so it is automated to the maximum. Unregulated and untested seems a bit strong to say the least. Or again, it can apply to any RWA.

To avoid any doubt, there is no such thing as a team. No one besides an MKR vote can make them do anything, and “them” is a regulated corporate servicer.

1 Like