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

Proposals
1

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

TL;DR

What this proposal does:

  • Give Arbitrum DAO more options for doing grants in stablecoins, making long term investments using stablecoins, and having safer access to capital without needing to sell treasury assets like ARB tokens.
  • Design and build a system for borrowing stablecoins that prevents liquidations via more assets being pulled from the treasury to increase collateralization.
  • Release that Treasury Backed Vault system publicly as a public good which can be used on any stablecoin protocol.
  • Complete a security audit on the smart contact code for the TBV, and a thorough risk assessment.
  • Report the findings and parameter recommendations to the DAO.

What this proposal does NOT do:

  • Give allowances to move ARB from the treasury to any CDP protocol.
  • Create debt or other positions with DAO funds.

Abstract

This pilot stage proposal seeks to complete Research & Development, as well as a comprehensive risk analysis, on a new Treasury Backed Vault (TBV) system which gives Arbitrum DAO the option to borrow stablecoins from existing protocols against ARB, with extremely reduced liquidation risk. It expands on existing discussions/designs of “DAO Owned Liquidations” for CDP protocols. Liquidations are always prevented by automatically pulling more collateral from the DAO treasury, increasing the collateral ratio of the loan.

Pilot Phase

This pilot is designed to give the DAO sufficient information to make an informed decision regarding borrowing stablecoins against assets in the treasury, and the confidence to use our proposed solution safely.

Motivation

The Arbitrum community is seeking ways to more efficiently and effectively use the capital in the treasury. The strategies for grant programs, Mergers and Acquisitions, long term investments, and treasury sustainability seek sustainable solutions that minimize required sell pressure on the ARB token.

The treasury currently has about $3.8B USD worth of ARB tokens. Long term diversification of this is key, but the community also needs access to capital now for grants, treasury management, M&A, and more. Developing an alternative strategy for accessing stablecoins gives Arbitrum DAO more options and power.

Key Terms

CDP - Collateralized Debt Position. A type of financial primitive used by protocol like Maker, Open Dollar, Reflexer, etc… Collateral backs stablecoins with required loan to value ratios.

Liquidation - an event which happens when the value of collateral backing a loan is less than the value of the loan, or below a required threshold. This forces the collateral to be sold in order to pay back the debt.

Specifications

The DAO should have the option to borrow stablecoins against its ARB and other future assets. However, using traditionally structured lending protocols creates too much risk of liquidation. Our proposed solution is a Treasury Backed Vault that severely limits the liquidation risks and lets the community achieve its goals with open access to its capital.

Liquidations are mitigated by a “savior module” connected to a CDP position. The savior module adds an additional step between a vault being marked for liquidation and becoming liquidated. The module automatically pulls more ARB tokens directly from the DAO treasury (or elsewhere), adds them to the collateral of the position until it is raised above the liquidation threshold, and completely prevents liquidations from occurring. Any time liquidation on the DAO’s position is called, the savior activates first. The module contract would be given an allowance of ARB tokens by the Treasury to spend, allowing the Arbitrum DAO to completely control and at any time update the bounds of how many tokens can be pulled from the treasury.

Example usage of Treasury Backed Vaults:

  1. Arbitrum DAO deposits 1,000,000 ARB, worth $1,500,000 in a CDP protocol like Open Dollar or others where Treasury Backed Vaults are enabled.
  2. Arbitrum DAO gives a TBV savior module permission to spend an additional 1,000,000 ARB, which remains in the DAO treasury. The DAO sets parameters with the savior.
  3. Arbitrum DAO borrows 1,000,000 stablecoins worth $1,000,000
  4. Due to market conditions, the price of ARB drops 33% suddenly, and the position is under the required collateralization ratio. It is now up for liquidation.
  5. Anyone calls the public function to liquidate the position. The vault is marked for liquidation.
  6. Before liquidation occurs, the savior module is automatically called. It has been set with specific parameters by the DAO to pull an additional 1,000,000 ARB from the treasury and deposit it in the vault.
  7. The position now has 2,000,000 ARB deposited in it, and is safely over the collateralization requirements. No liquidation happens and no additional votes were required to manage the position safely.

Steps to Implement

The implementation is essentially split into 4 phases

  1. Research and Development

A team of 3 solidity developers design, build, and document Treasury Backed Vaults with support for Arbitrum DAO specifically in mind.

  1. Risk Analysis

A team of 2 financial analysts complete a report on suggested parameters and the greater ecosystem impact of different borrowing scenarios for Arbitrum DAO. This will come with many helpful recommendations that future delegates can lean on to make proposals to use the vaults how they choose.

  1. Audit

A public audit contest will be conducted and managed by our team to review the security of the smart contracts of the TBV. This is important to complete at this stage because we want the DAO to be able to feel confident in the readiness of the system. A formal security report will be published and summarized in a post to the DAO. Audit reward recipients will be required to complete compliance and KYC before receiving funds.

  1. Reporting

At the end of the pilot a report will be made to the DAO that includes results of the security audit, documentation on using TBVs, and results of the risk analysis. The report will include recommendations like how much ARB should be allowed to be pulled by a TBV to limit risk, how many stablecoins could safely be borrowed at a time if the DAO should choose to use the TBVs, how much sell pressure on ARB could be avoided in different scenarios, suggested collateralization settings, and other parameters for real usage.

Timeline

Overall time to project completion: ~2.5 months

Research and Development, done asynchronous with risk analysis reporting: 1.5 months

Security audit and reporting: 2 weeks

Complete report for the DAO and documentation on usage: 2 weeks

Team

Research and development will be completed by a designated team made up on contributors to Open Dollar and led specifically by CupOjoseph & Pi0neerpat. Open Dollar is the largest Arbitrum native stablecoin protocol by TVL. We will also oversee the third party audit of the code that is developed, and personally present the risk analysis (including recommendations) to the DAO at the end of the pilot.

Overall Cost -

Total cost: 250,000 ARB

Breakdown

Research and Development: 125,000 ARB

Independent Security Audit: 50,000 ARB

Risk Analysis Report: 50,000 ARB

Administrative: 25,000 ARB

Remaining or unspent funds will be returned to the DAO treasury promptly.

Thank you!! Please support our proposal, we think this could be a huge unlock for the DAO and a powerful tool to add to the community’s toolbox. We appreciate any further feedback on and off the forum on this proposal and the Treasury Backed Vaults concept overall.

Helpful links:

Treasury Backed Vaults Overview from Open Dollar

Original DAO Owned Liquidations post

Arbitrum Treasury Sustainability Group discussion on DAO Owned Liquidations

Existing proof of concept implementation. That’s right, we already started building. :sunglasses:

EDIT:
Based on the feedback so far we will publish this week a list of specific research and modeling questions that should be answered in the risk analysis report.

Risk Analysis Report Topics

A list of topics, as requested by the community, that will be covered in the report. Each will have a model of different scenarios and specific recommendations to take on short, medium, and long term time horizons.

  • Over-Collateralization Ratio requirements for loans are different sizes
  • Interest rate settings for loans are different sizes
  • Loan sizes based on different market conditions (including worst case scenario)
  • Propose Debt ceiling requirements on CDP protocols that integrate TBV
  • Sell pressure situations on the stablecoin
  • Strategies for paying down debt long term given different market situations
  • Percentage of the stablecoin which is backed by TBV or ARB token versus baskets of other assets
  • Outline Mechanism for reporting on the health of TBV positions
5 Likes
2

This is an interesting idea, but what would be the natural way to unwind the debt? After the CRV near-miss, protocols are unlikely to just let Arbitrum pledge near-infinite amounts of ARB to continue collateralizing the debt. What’s the protocol to reduce the debt and not just increase the collateralization? Or are we assuming funds will become available from the STEP or other programs to service the debt?

2 Likes
3

Thank you for the thoughtful reply, @GFXlabs! I really appreciate your time.

Excellent question.

  1. As the DAO makes investments, this will theoretically drive value to ARB token. Right now treasury shares in less and less of the upside as we make investments through essentially selling ARB. With more ability to borrow against ARB the treasury will share in more upside, and be able to sell at higher prices in the future.
  2. The DAO can sell ARB at any time and pay back positions
  3. Specific to Open Dollar, CDP positions are stored as tradable NFTs called Non-Fungible Vaults, so the DAO could sell this entire position to a market making partner like wintermute at any time to close it simply.

IMO this is the largest current blocker to the DAO being able to borrow significant amounts of stables. By lowering the risk of needing to liquidate a large position, a TBV system should make the DAO an even MORE appealing customer for CDP protocols. Part of the risk analysis section will include recommended collateralization requirements and debt limits for CDP protocols who decide to implement the TBV.

This proposal makes no specific strategy recommendations, deploys no funds, and creates no debt. It simply gives the DAO a new option to create debt if the community wants to in the future. IMO, options are good! We will work closely with treasury management teams to ensure TBV usage makes sense long term.

Hope this answers your questions! I would love to chat more any time.

1 Like
4

This is an interesting proposal that may solve some important problems. I am especially enthused by the saviour module.

A little further down the line, I am wondering what your approach to the sell pressure on this dollar will be? Grant recipients (that is the primary demographic here, correct?) will need to cover operating costs and will have to swap into assets that can be offboarded from DeFi.

GHO experienced difficulty with its peg and has settled at $1 after Aave DAO directed significant resources at it. It is indicative of the complexity synthetic stablecoins face that this new treasury backed stablecoin would potentially deal with.

3 Likes
5

This is a great question. We can add this to the risk analysis report specifically. It will definitely include recommendations to CDP protocols on how much debt to take, what % should safely come from the DAO, etc…

Borrowing a stablecoin without deep liquidity is much less useful than giving the DAO the option to more safely plug into existing ecosystem on Arbitrum.

To be clear, we are not suggesting creating a new stablecoin. But rather, plug in the savior module to existing stablecoins on Arbitrum like Open Dollar (but in an agnostic way so the DAO is not picking winners). If the stablecoin is backed by a basket of assets and arbitrum DAO is just one of the users we dont have to bootstrap liquidity from the ground and there will be enough market participants to keep the stablecoin pegged safely.

1 Like
6

I like this proposal for a couple of reasons:

  • It creates an avenue for the DAO to directly use Arbitrum protocols, which I think is a win-win.
  • It creates an avenue for the DAO to access stablecoins, which I think is important for DAO ops and treasury diversification.

Some potential concerns to consider:

  • The DAO will need to have an alternative source of stables available to pay off the debt per the concern of @GFXlabs. I don’t think the DAO can sell ARB for stables fast enough via governance for this use case.
  • Moving DAO ARB into DeFi protcols introduces additional risk to governance. I’d like to have a robust plan for mitigating risk (cap on total ARB deposited, plan for gradually increasing TVL, audits, etc.)
2 Likes
7

Thanks for the support Frisson!!

These are good questions and the additional risk that lending creates is a valid concern. That’s why such a big part of this is doing risk analysis on top of everything that gets built. We need to do this as a pilot first and make decisions on how to use the financial tools in front of us carefully.

Existing CDP systems are certainly too risky IMO for DAOs like arbitrum to use effectively. If collateral prices go down they require active management. That’s why TBVs can be such a game changer: no active management, just set an allowance once.

Future proposals can guarantee before depositing that the ARB is not being delegated and can’t be used in governance attacks. For example, Open Dollar currently delegates ARB deposited to

Future proposals could delegate some sequencer revenue to auto paying down debt without DAO management.

ARB tokens always require approvals to move, so theres no way a TBV could ever move more tokens than the DAO has explicitly voted should be able to move.

All of these suggestions would have recommendations in the final report. Thanks for this feedback and let us know if there are other risks we should focus on. We will take any other community suggestions what the community whats to learn and what hypothetical situations to model out.

On audits, etc:
I think we should fund multiple audits, like 5 or more, before actually putting millions into any defi protocols. But we want to keep our proposal lean enough to have a great ROI. No reason to spend $1m on audits before the DAO has seen risk analysis recommendations and decided whether or not the community even wants to pursue this option.

8

I think this is an interesting idea. A couple questions that come to mind are:

  1. What are the limitations of the savior module? If $ARB is in a death spiral, is there a point at which the savior model no longer adds additional collateral?
  2. If there is a point wt which the savior module stops, will the position be liquidated?

Savior module sounds great from the arb treasury standpoint, but it could potentially have LUNA like symptoms due if $ARB were to become extremely volatile.

2 Likes
9

Thanks for the question and feedback :pray:

  1. The main limiter in that case would be the allowance set by the DAO on the max number of ARB that can be pulled. Technically Arbitrum DAO can mint new tokens if I recall correctly. To keep the stablecoin backed a liquidation would have to be processed at some point.

  2. Yes, if the DAO abandons the position and no more ARB is pulled, normal liquidations will happen according to the protocol. (According to the current design)

This is not like Luna, since ARB is not the backstop or governance token for any stablecoins. We intend to make this available for existing stablecoin protocols so stables are backed by a basket of assets and not dependent solely on arbitrum DAO for stability.

For example Open Dollar has a debt limit of 5M stablecoins that can be minted against ARB, but another 20M that can be backed by rETH and wstETH
Thanks for the feedback, if you have other suggestions on how to communicate this difference clearly I’m all ears.

2 Likes
10

why triggering the saviour module at liquidation threshold instead of highening it at, i don’t know, 1.10 health factor, and bring enough funds in it to send it back to a target of safety which could be 1.2, 1.3 or whatever?
AKA why not doing it incrementally, and avoid having to act during a liquidation?

1 Like
11

This is a great idea, we can try to incorporate this! thanks for this feedback.

The main point of doing it when liquidations are marked is that these hooks already exist in many CDP protocols (like reflexer, Open Dollar, HAI, etc…) and somewhat similar savior systems. So we can build this in a way that it can sit on top of these existing stablecoins without requiring us to launch our own stable from scratch or make any changes to protocols that are already deployed.

Functionally I think its pretty similar tho, and if the price is fluctuating a lot you only want to pull from the treasury as little as possible IMO, so only when really needed.

12

Hi @cupojoseph,

Thanks for the proposal! It’s certainly very interesting and a novel way for the DAO to use its ARB to unlock liquidity via stablecoins. I wanted to clarify a few questions about the process:

  • In a situation where ARB price decreases consistently and the savior module is full, are there any mechanisms to avoid a liquidation cascade if the entire position gets liquidated and there is an even larger price crash? Processing liquidations in this scenario could lead to a death spiral, so price pressure may need to be minimised with very risk-averse debt ceilings.
  • Moreover, the debt ceiling for ARB may need to be pretty low, given the low levels of on-chain liquidity for ARB (as mentioned by Karpatkey and Gauntlet in their research reports). In such a situation, would there be scalability issues with employing this solution for larger initiatives, i.e., would using TBV’s be limited to smaller grant programs?
  • What assets will the DAO use to pay the CDP back? As Frisson pointed out, it probably cannot be ARB because even that will add sell pressure (however minimal), so what would you suggest?

Overall, would love to see these issues ironed out and a practical solution to be found as we agree with the premise behind the TBV: seeking sustainable solutions that minimise sell pressure on ARB.

1 Like
13

Hey everyone, the Treasury Backed Vault (TBV) proposal sounds useful- especially that “savior module” idea! I had a couple of thoughts:

  1. Keeping it transparent: The savior module has a lot of control over the DAO’s money. How can we make sure everyone knows exactly when it kicks in and how much it can grab from the treasury? Maybe a clear system for community votes on those settings?
  2. While TBVs seem like a great solution, have we considered other ways to manage liquidation risk, like interest rates that adjust based on market conditions or diversifying the DAO’s collateral portfolio? It’d be good to consider this imo.
1 Like
14

Awesome questions Sid, thank you so much for the close revue. Areta’s M&A strategy proposal that was just approved is a big inspiration for this proposal.

  1. What do you mean by “module is full” exactly? If theres no more ARB allowed to be pulled from the treasury normal liquidations can happen. Our liquidation engine is essentially the same as Maker with a few efficiency improvements. Unique to Open Dollar, all CDP positions are stored as NFTs that themselves can be sold instead of liquidated. It’s possible the position could be handed off to a market maker partner like wintermute in that scenario instead to wind down. As opposed to market sold into spot markets during a normal liquidation

  2. agreed, to prevent #1 from happening and avoiding spirals there needs to be conservative debt ceilings. For example, Open Dollar debt ceiling for ARB is only 5M OD, while rETH and wstETH are each 10M. We want to build this so it can sit on top of existing protocols, and we’re only backing a portion, (maybe 20% max?) Ultimately the investment helps us do a lot of modeling of different scenarios. TBV could be for long term investments that are better made via stables.

  3. If the DAO needs to pay back $10M in stablecoin debt which it issued for some program, they would have otherwise have to sell $10M ARB right away to access that capital. Ie theres no new sell pressure. Long term I think there will be staked ARB generating rewards, sequencer revenue shared with the DAO, a stablecoin L3 that pays fees back to the DAO treasury to support paying back debt, diversification into interest bearing assets that pay down debt, etc… lots of long term strategies that should be considered individually & separately IMO.

Hope this answers the questions a bit, and I think it will be really important to spend significant time on modeling different price scenarios and debt ceiling requirements so the community can make the best informed decisions.

15

Thanks for the review Juanrah! Glad you see how useful this could be.

The TBV can not pull any arbitrary amount. The DAO must call the approve(TBV Savior, amount of ARB) on the ARB token contract to give approvals. So it could only pull a maximum amount that was explicitly set. And that amount could be increased or decreased to certain bounds by the community at any time.

Transparency is a top priority. It should be designed so there are never any surprises, and the amount of allowance the TBV has at any given time is always clearly community.

:heart: :heart: :heart:

If you have something specific in mind I would love to add it to this proposal or otherwise include it in the research. Ideally we are building this to plug into existing CDP protocols without having to make any changes to them.

Borrowing against ARB would give the DAO capital to diversify into other assets like GMX pool tokens that bear yield, LSTs like lido or rocketpool ETH, etc., without having to create tons of sell pressure on ARB. I actually would not support that personally, because its too much leverage/risk, but it would be a new option for diversification long term if the community wants it. Having the TBV decreases risk of such actions.

1 Like
16

Thanks for putting in the work on this! Its gratifying to see research from our treasury working group converted into a proposal and hopefully a product very soon.

@sids2000 research also proposed the reverse of a saviour module, where if the ARB price increases a lot it is sold to pay down earlier debts.

However, I don’t see this aspect referenced in the current proposal. Is there thoughts of including a reverse of a savior module , where ARB price increases beyond a threshold pull in ARB to pay down debts?

1 Like
17

Sids2000 research was the original inspiration for this project and proposal completely. But we left this feature out for these (reluctant) reasons:

  • The DAO can already pull funds out in a bull scenario when ever it wants.
  • Automating selling ARB is a very big decision not to be taken lightly by a bot IMO. In the future if ARB is $1,000 each and our system automatically sells millions of dollars in ARB token without anyone in the DAO voting for that to explicitly happen, I imagine that would be pretty upsetting.
1 Like
18

This proposal is very interesting, more efficient treasury management for Arbitrum DAO by enabling stablecoin borrowing against ARB assets without liquidation risks. It enhances financial flexibility for grants, investments, and operations, ensuring access to capital while preserving ARB value. The research, development, and audit phases ensure robust, secure implementation, ultimately benefiting the DAO with diversified funding options.

19

Thanks for this interesting proposal @cupojoseph

  • Will your research include, an analysis of the long-term implications of using TBVs, including potential scenarios where the savior module might need to activate frequently due to prolonged market downturns?

  • While the proposal mentions a security audit, it would be beneficial to outline a continuous security & risk management plan, including periodic audits, bug bounties & real-time monitoring.

  • The proposal could benefit from outlining specific mechanisms for regular reporting, community engagement & feedback

1 Like
20

Thanks for feedback Jenga!

All of the research topics mentioned so far will have short, medium, long term modeling for different market scenarios. I have updated the proposal to make this more explicit in the new Risk Analysis Report Topics section.

A bug bounty is a great idea, and we can potentially work with Pessimistic which has gotten ARB grants to do audits for Arbitrum protocols before. We want this to be production ready with extremely high confidence in security at the end of the timeline for this grant.

IMO continuous security monitoring and regular reporting should 100% be included and even required in any future proposals to actually deploy funds with a TBV.