[Non-Constitutional] Pilot Stage: Treasury Backed Vaults research and development

This proposal is way more straightforward than I initially thought, I think it would be great to confirm that there is buy-in from a few other projects in the ecosystem to use this for their own treasury moves, (even if it’s just with rETH) before throwing down $250k into the research.

At first thought, it doesn’t seem like a good idea for most DAO initiatives, but it does seems suitable for some, specifically investment-style proposals with expected returns or proposals that explicitly have an objective to make ARB number go up. However, there should be stop losses and plans to pay down the loans at specific ARB price targets as part of any proposal that wants to integrate this very cool tool.


Thanks for the proposal @cupojoseph, this is an interesting idea, however, we do have some reservations about whether or not this is something the DAO needs as stated in its current form.

A mechanism like this seems good because the DAO refrains from selling ARB to fund grants, expenses, and investments. However, assuming most of the borrowed stablecoins are used to cover expenses (or grants/investments) at a modest LTV ratio we end up with more and more ARB locked up in the CDP as the price goes down or expenses go up (or both). So we are locking a large amount of ARB to access a small to moderate amount of stablecoins to cover expenses and this position remains locked until the debt is repaid which will either require the DAO to sell ARB (which is bad for the position) or finance it through other alternatives such as the current treasury diversification initiatives or sequencer revenue. We’d much prefer the DAO to continue focusing on treasury diversification efforts that are more capital efficient which can then be used to fund the DAO’s expenses.


Thanks for this post, as it is a very interesting topic and solution proposed.

My 2 cents in here are just cost-related: Is it possible to leverage the ARDC structure (namely the risk management & security members) in part of this research?

Tagging @Sinkas for visibility.


Thanks for the review Griff! We have had several internal conversations with defi protocols that are interested in this system, but want to see what the final implementation will be before commiting to support. I’m trying to round up some more support more publicly from them now and expect some protocol teams to vote in the snapshot.

Thanks for the review team!

Is there something else you think is missing from the research that we can add or focus on in our implementation to meet your concerns?

What is the borrowed stables are used for investments with a specific plan to pay down the debt and an expected EV higher than any fees paid on the loan? personally, I dont see investments, especially long term ones, as expenses. Selling ARB to make investments leaves tons of upside on the table. A 1% change in the price of ARB = $30M loss or gain for the DAO treasury right now.

This is not necessarily the case. With the Open Dollar CDP protocol for example, debt positions are tradable and can be sold OTC without needing to pay back the debt.

I think we can agree that treasury diversification is the end game here for sure. Wintermute can diversify its fund easily because it can take uncollateralized loans any time. Arbitrum DAO has to be more creative. If you dont support deploying capital this way, please consider supportting the research we are doing to come up with new strategies for the DAO.

1 Like

Definitely! If they are open to supporting. We will take all the feedback and support we can get.

I am not opposed to more research being conducted on treasury backed vaults, but I would like the end goal to be an RFP process rather than funding an entity to build TBVs from scratch.

For example, I believe Curve’s crvUSD that introduced soft liquidations via LLAMMA pools would be highly compelling for the use case being described. We even granted Curve ~240k ARB a few weeks ago to create a market for ARB/crvUSD/CRV and soon LRT/WETH pools. Why would we not want Curve (and others) to be able to throw their hat in the ring for this type of initiative?

I also think the price tag is too high here - I think 100k ARB is more inline with the proposed outputs. We have the ARDC for a reason, and if DAO treasury utilization for borrows is something that is desired, we should utilize the resources we have already paid for to conduct the R&D. An added benefit of utilizing the ARDC would be the lack of bias in the report.

I have reservations as a whole on the proposed TBVs, as a vast majority of the DAO’s spend is not expected to be returned. In the cases where we DO expect a return (see the GCP), the value is so large that a TBV would be rendered entirely infeasible anyways. I understand people here are bullish ARB and want to avoid selling it, but does this actually make sense in practice? My gut tells me no…

1 Like

I was a bit surprised when this popped up on snapshot today. This proposal was brought to life just 1 week ago in the forum, and there are several unanswered questions from several folks.

First, I’ll report my opinion, already shared privately, and then a considerations on the proposal.

I see 2 scenarios here:

  1. we create a vault for which we establish a cap. Let’s say is 50M stables. And we back it up with 500m arb between collateral and reserve, meaning arb has to go to 0.10$ to get liquidated. We have something that let us get $50m in liquidity, but this amount is basically 1 incentive program for us. All of this by using half a billion arb. Feels like the amount received is too small to have an impact, and so is not that important considering we would have to put on stake half a billy of arb to do so.
  2. we decide to use the vault, with a certain cap tied to the arb price, over time. Which means, while w allocate 300-400m arb between collateral and reserve, we found ourself with arb going to 10$. We now a 4B collateral we can borrow against. And we still borrow 1:10 ratio so we borrow 400m. That is a meaningful amount.

Let’s not consider scenario 1, because to me scenario 1 is not likely important nor worth the effort. Let’s only consider scenario 2, with whatever degree of numbers you want (amount of arb, price of arb, amount we borrow), keeping in mind tho that we want numbers that have an impact, so we are on the high end of the barbell.

Whatever amount you have as collateral. Whatever amount you have in the reserve. Whatever amount you borrow. You now have a target on your back.
You have a series of number which, summed with depth of liquidity of dexes and cexes, can be put in a spreadsheet to calculate how much capital you need to liquidatate that position.
This could be economically viable, or not. But the moment in which it becomes viable, trust me someone will exploit it.

I want to state that

  1. this is obviously my personal view
  2. I am being pessimistic here
  3. I am also realistic tho.

And this was my take on the product. I also reflected upon it, chatting with proposers, and agreed that, for sure, we can do researches on this.

Which takes me to the second big surprise that I have: why are we packing the research with the product? Shouldn’t we first just research if this is viable, or if we are just falling in a crv type of scenario, with the difference that mitch can pick up the phone, his trezor, send money around, gather collateral and manage risk and we, as a dao, are as slow as a crippled elephant?

First, thank you for bringing this proposal. It’s one of the more interesting ones mechanically.

After giving it some thought, we’re not opposed to the concept, but would prefer governance avoid borrowing against ARB as a collateral asset. This boils down to two reasons:

First, ARB in the treasury should not be considered an asset. From an accounting standpoint, counting that uncirculated ARB as an asset would violate accounting norms and perhaps even be alleged to be a misrepresentation. GFX believes in firmly stamping down the misconception that uncirculated native tokens are assets. It encourages excess spending due to a feeling of wealth, and is simply incorrect from an accounting standpoint. So utilizing ARB tokens as collateral would – aside from other objections raised by others – be the same as minting new ARB if a liquidation were to occur.

Secondly, and more optimistically, Arbitrum has other assets it could leverage for credit. Governance has access to ETH, and soon will have stable-value assets as well (even if they’re more exotic or legally encumbered RWAs). To the extent that credit is needed, we would prefer to see existing assets used as collateral instead of a dilutive mechanism.

More broadly, slowing down dilution from spending ARB by making new spending a mixture of ARB and borowings is appealing. We support continued exploration of this kind of financial planning. but we do not support any significant spending or moving forward with such an arrangement until a clear need to borrow is identified and a plan to repay the debt is included.

Attaching borrowing to a smaller upcoming initiative would be a way to test out governance debt as a concept, and see if credit can be responsibly managed by governance.

1 Like

The most important question is that I really don’t understand why the Arbitrum needs this.

  1. When borrowing stablecoins, ArbitrumDAO will have to pay interest. Considering that ArbitrumDAO does not need stablecoins in its operations, it will also lose money on interest.
  2. I do not see any effective use of capital in this proposal; on the contrary, there will be losses.
  3. Nobody bothers ArbitrumDAO to put ARB as a 10/1 collateral and there will be no liquidation by default

We’ve already put a lot of work into preparing this proposal and the feedback we’ve gotten from risk analysis is that it is much easier to begin a study on the risks if we have a given implementation and spec to go off of. It’s harder to judge the risk of a system without an actual spec of it to go off of. I think thats one of the reasons that the original DAO Owned Liquidations proposal has taken this long to move forward meaningfully.

We are that slow right now. Which is why actually using our treasury capital in DeFi safely is impossible. I see this grant as a giant step towards automating risk management on-chain and giving the DAO more tools to do so.

Agreed. The system is not specific to ARB token, and other DAO treasuries on arbitrum can make use of it as well. ETH if the Arbitrum treasury had tons more ETH, it still wouldnt be in a position to use that in CDPs safely without many extra safety rails, like TBVs.

Thanks for this thoughtful feedback GFX! I think there could be a day in the near or medium future where Arbitrum needs capital quickly but, because its not a traditional business, getting a credit line is impossible. We need to set the groundwork now and have high confidence in a CDP system that can be deployed quickly (days not quarters). Even if we dont aim to deploy any capital in the near term.

If you have one in mind, I’m definitely open to discussing!!

Arbitrum needs this because the DAO is incapable of accessing cheap lines of credit. Also, if our only strategy for spending on grants and making long term investments is to sell ARB, then we are leaving all the upside on the table.

Big assumption here IMO. Selling ARB forever is not a sustainable growth strategy.

Thanks for tagging. Is the thought to use ARCD if the proposal passes, to help with the research, or to leverage the ARDC before the proposal passes to potentially help improve it?

Ideally both :wink:

But my question was regarding lower the proposal cost: if it is possible to use the ARDC structure to conduct part of the research.

1 Like

Would be dependent on several factors, but potentially yes

1 Like

I would hope they want to work with us, but the members of that council also only have so much bandwidth and the risk analysis portion of this pilot is pretty significant amount of work for a whole team. We have taken feedback to get this far from several members of ARCD and also the arbitrum treasury sustainability group, who original proposed the idea of DAO owned liquidations. But I cant speak for any ARCD members

This proposal is necessary which outlines a strategic approach for the Arbitrum DAO to borrow stablecoins against its ARB assets through a new Treasury Backed Vault (TBV) system, significantly reducing the risk of liquidations. The plan includes research and development, a comprehensive risk analysis, and a security audit, ensuring that the DAO can make informed decisions about using stablecoins. By implementing the TBV system, the DAO gains more flexibility in capital management without selling ARB tokens, which is crucial for funding grants, investments, and other strategic initiatives. The proposal is well-structured, aiming to enhance financial efficiency and security for the Arbitrum community. I fully support this initiative as it provides innovative solutions for sustainable treasury management and strengthens the DAO’s financial infrastructure.

On behalf of the Arbitrum community members who delegated their voting power to us, we’re voting Against this proposal in its current form. While the concept of Treasury Backed Vaults is intriguing, we have concerns about using ARB tokens as collateral. We would prefer to see a smaller pilot tied to a specific initiative with a corresponding budget, along with a clear plan for debt repayment, before pursuing this at scale. It should be noted here that the Arbitrum DAO already has several promising initiatives underway to productively deploy treasury capital.

Hey @mcfly thanks for the review and engagement! I want to address your points briefly:

This proposal is meant to develop a system for Arbitrum and other DAOs. It does not need to be specific to ARB. Even if the Treasury was full of ETH, with our current set of tools we would be unable to borrow against it or use it efficiently.

Since this proposal deploys $0 and creates no positions, it’s hard to imagine what a smaller pilot would look like. If you have other ideas on how to limit it further please let us know! We want to improve the proposal as much as possible.

This is something that arbitrum DAO, and specifically the treasury sustainability group, has already put resources into proposing, but has been unable to actually move forward with any implementation until now. We feel it’s essential for our long-term projects that you mentioned that we research new ways for the DAO to activate capital once it is diversified.

I’d love to discuss one-on-one if you’re open to it. And we’re completely open to limiting the scope if you have ideas for what you mean specifically by a smaller pilot.