GIP-86: Should the GnosisDAO commit to additional rETH usage in treasury operations?

[Viszla]

GIP-86: Should the GnosisDAO commit to additional rETH usage in treasury operations?

• In Favour
• Against

GIP: 86
title: Commit to Additional rETH Usage in Treasury Operations
author: AnotherViszla
status: Draft
type: Meta
created: 2023-05-10

Category

Governance/ Treasury

Executive Summary

Gnosis DAO / Chain has decentralization as its core ethos, focusing on a distributed validator set for the chain and an open DAO structure run by GNO holders. Despite this, our Treasury has nearly all of its staked ETH assets sitting with LIDO as stETH. Given the success of RocketPool Protocol in creating a massively decentralized validator set through rETH and RPL, this proposal recommends that Gnosis DAO commit to moving at least 50% of their stETH holdings to rETH holdings over time in order to support Ethereum decentralization and reduce single protocol and “bad actor” risk that comes from Lido’s limited node set (29 nodes vs Rocket Pool’s 2550 and growing nodes).

Introduction

As of last notice, the Gnosis Treasury (managed by Karpatkey) holds 64,124 wstETH (~$130mm) and 5,725 stETH (~$10mm), 712 rETH (~$1.4mm), and 18,622 ETH. Full report here: 20230424 - karpatkey’s Weekly Report on GnosisDAO Farms.pdf - Google Drive This concentrated holding of a centralized staking provider token is a risk to Gnosis DAO and against the DAO’s core ethos of decentralization. This GIP is intended to direct Karpatkey to move to reduce stETH dependency and introduce additional rETH into their DeFi strategies in order to reduce risk to the DAO and provide support to decentralized ethereum staking.

Proposed Changes

The Gnosis DAO will instruct Karpatkey to reduce stETH dependency and increase staked ETH diversity, but WILL NOT dictate how this is best accomplished. Karpatkey, once instructed, will move forward with whatever strategies are in the best interest of the DAO to reduce stETH exposure and provide additional usage of rETH / RocketPool

Impact and Benefits

Over 500,000 ETH is currently staked with Rocket Pool, making it the third-largest TVL after Lido and Coinbase among liquid staking providers.

Diversification into rETH is wise for a number of reasons:

1. Insurance

rETH has best-in-class resistance to black swan events, such as a massive slashing resulting from client bugs.

If a slashing event were to occur, funds are first deducted from the following sources before impacting rETH holders:

1. Node operators’ ETH Share(~185k ETH)
2. Node operators’ RPL Bonds(~200k ETH)

In contrast, Lido’s insurance fund only has around 6k stETH total, which (A) is stETH-denominated itself and (B) services 25 times more TVL.

2. Peg Performance

rETH has shown more resistance to negative de-peg than stETH.

Source: Dune Analytics

Both stETH and rETH are expected to be better able to maintain their respective pegs via significantly different methods. Lido will request their NOs to exit as needed to restore supply/demand balance. Rocket Pool will primarily balance using deep liquidity and market arbitrage opportunities for rETH holders. If needed, arbitrage opportunities for NOs that create/exit validators serve as guardrails for the peg.

3. Tail Risks

Every liquid staking protocol has some level of risk.

Rocket Pool’s decentralized and on-chain nature means that protocol-related bugs are the primary concern. Comprehensive and continuous audits, bug bounties, and an engaged community of node operators are the primary methods of mitigation.

Lido’s group of trusted node operators with no personal bond currently means that a node operator could keep all execution layer rewards for themselves and refuse to exit validators. This risk is mitigated through legal contracts and by specifically choosing operators that are trusted members of the Ethereum community.

While one can argue which tail risks are larger, a key practice in risk management is diversification.

Implementation Plan

Once support is indicated in the above poll, we will move this GIP to a formal on chain vote. The implementation will simply consist of instructing Karpatkey to begin transitioning our Treasury away from stETH and towards rETH. The expectation is that Karpatkey will work to reduce slippage associated by any conversions by swapping stETH for rETH in smaller batches, or in some cases waiting for Lido redemptions to unlock if swapping on DEXs is not feasible.

Risks and Challenges

1. wstETH is currently being used by Gnosis DAO to collateralize our Maker loans of ~22mm DAI. Any actions to reduce this and replace the collateral with rETH will need to be carefully managed by Karpatkey to prevent any liquidation risk.
2. Assuming Gnosis DAO wants to move some stETH exposure to rETH ahead of the Lido withdrawal unlock, Karpatkey will need to manage any slippage associated with the swaps. Today, the stETH discount is just 0.4%, but my recommendation is that we move the majority of our stETH position to rETH once withdrawals are enabled. 0.4% is significant when considering the size of the Treasury and Lido unlock is expected around May 15th. Given this poll closes on May 15th, this may be a moot point

Team/ Organization/References

Viszla is not a team, just a poorly spelled dog reference. I am a solo ETH and GNO staker, hold RPL, and plan to stake all future ETH with the RocketPool protocol.

Gnosis Snapshot

TBA

[DrDoofus]

I should add, there are a lot of GNO token holders, Gnosis builders, validators, etc. from Rocket Pool who are joining to vote. I am a gnosis builder, GNO hodler and Rocket Pool node operator and I think this is a great way for Gnosis to show support for decentralization while gaining security (and value) for the treasury.

Now that Viszla has shown us the way, I am reminded to come here more often and exercise my GNO vote!

[Karpatkey]

Though the ethos that supports this proposal is widely shared across the community, this proposal may set a dangerous precedent. Optimising exclusively for decentralisation would mean acquiring risk from multiple protocols that might not be as battle-tested, audited or liquid enough to manage.

From a governance perspective, this proposal also conflicts with GIP-58, where karpatkey is granted autonomy and tasked with the responsibility to design overarching strategies that drive treasury risk-adjusted growth and support the DAO’s goals, including the Gnosis Chain development. Having an individual proposal that partially affects the non-custodial treasury management would also misalign the incentives and objectives of the Gnosis-karpatkey partnership, setting the wrong precedence moving forward.

Rocket Pool’s project is noted, and in fact, the DAO holds rETH and provides liquidity in Balancer. We will continue to monitor its performance, security and market conditions for opportunities to reduce risk, support the Ethereum ecosystem, and increase the treasury’s growth (currently, there are 4 rETH for 100 (w)stETH in the market - source: Nansen).

New ideas or perspectives are always welcome, and we encourage you to share them on the treasury forum thread or reach out to us directly through Twitter.

[DrDoofus]

Interesting. I suppose if we don’t optimize for decentralization, we should not be surprised when we don’t get it.

[Viszla]

I agree with these points, which is why I was careful to include that “The Gnosis DAO will instruct Karpatkey to reduce stETH dependency and increase staked ETH diversity, but WILL NOT dictate how this is best accomplished. Karpatkey, once instructed, will move forward with whatever strategies are in the best interest of the DAO to reduce stETH exposure and provide additional usage of rETH / RocketPool”

Karpatkey still has absolute control over the Treasury strategy, but with this vote the DAO is signaling that our point risk to LIDO’s small validator set is dangerous for the DAO. We have already seen one of their validators get slashed and are concerned by LIDO protocols small protections against large failures. The execution of a plan to reduce exposure and increase decentralization would be entirely Karpatkey’s. This is similar to how the DAO can elect to spend as much of the Treasury fund as is confirmed by token holders, but Karpatkey decides how best to provide the $ to be spent.

[Karpatkey]

Thanks @Viszla. We appreciate your input and welcome ideas, especially those supporting decentralisation and other initiatives to increase our ecosystem’s long-term sustainability.

GnosisDAO already has a position in Rocket Pool. We’ll consider additional strategies using rETH and other LSD tokens in the future, granted they comply with the DAO’s risk profile and provide strategic value to the treasury and the community.

[Viszla]

Would welcome anyone else’s opinions here before this is taken to an on-chain vote. Is there opposition to signaling Gnosis DAO’s intent to reduce stETH exposure and add additional rETH to our LST portfolio? I understand Karpatkey’s hesitancy to allow any fine-tuning of their Treasury strategy, but i believe a broad vote to favor decentralized LSTs over time does not trample on their ability to execute that mandate as they see fit.

I would note that presently, the Treasury has USD137mm equivalent stETH and USD1.9mm rETH. This compares to relative protocol sizes of 6.3mm ETH (lido) vs 592k ETH (rocketpool).

[refri]

I am clearly in favour of diversify the LST and shift some more in direction of rETH. But ofc I am not educated in risk management and as you already said: This proposal isn’t done to override the work of @Karpatkey. If there are valid arguments against this I would be happy to know, even if it’s only the market cap as said above. Instead of arguing with conflicts with GIP-58 I would appreciate a more in depth explanation on this special topic, although I totally understand that this can not be expected for every investment decision.

[PhiMarHal]

I found this topic a couple days ago and was surprised to read @Karpatkey 's answer.

Today, saw context that seems relevant:

https://old.reddit.com/r/ethfinance/comments/1412u9x/daily_general_discussion_june_5_2023/jmyt7az/

Karpatkey also manages Lido’s treasury. Is there a conflict of interest here?

[adcv]

Hello,

I’m adcv from Steakhouse Financial, who operate the Finance workstream on behalf of Lido DAO. I helped write the original Treasury Management Committee proposal, which we seeded with Karpatkey because in our view they are the best-bar-none in DeFi treasury management.

If you or other commenters you have linked to had taken the time to read it before posting, you would have found that it is incorrect to say that “Karpatkey is managing Lido’s treasury”. Karpatkey is one of 7 entities on a multisig designed to trigger Aragon motions or EasyTrack calls for a very restricted scope of actions and in no instance ‘manage’ or take custody of any funds from Lido treasury. It’s very different to any of their mandates with DAO treasuries.

As an example, the first proposal this Committee has put forward is to simply stake all of Lido’s ETH. In the future, the Committee may propose other ideas and token holders may propose to change its constituents, overrule the Committee or launch Aragon motions directly. The ultimate aim of this Committee is to write very simple automatic rules and then disband.

The results of Karpatkey’s tireless work ethic do not need insistence or engagement amplification, they are plainly visible to everyone on the blockchain. The DAOs that continue to entrust them with designing non-custodial strategies are themselves a testament to the staying power of Karpatkey’s work. There is no conflict of interest, plainly, largely because none of their mandates are a secret and their only incentive is to work in the interests of the endowments and treasuries that they help manage. In fact, as they themselves attest above, it would seem that they independently arrived at the conclusion that some rETH allocation was appropriate for Gnosis already, before this proposal.

Regarding the merits of this proposal, I personally believe Gnosis DAO token holders should be able to make a considered decision based on cold scientific merits, without concern trolling or consensus cracking from community members. Ideally, all discussion on this thread ought to be about the strength of various ideas, not about the intensity of people’s emotions. It is not in Steakhouse’s policies to try and interfere in DAOs we do not contribute to and it is not in Lido DAO’s culture to astroturf support for stETH. There are good reasons for and against allocation choices to various tokens. The best feature of this or similar proposals is that no matter how it turns out Gnosis will continue maintaining an allocation to decentralized ETH liquid staking tokens, which is itself an important goal.

[PhiMarHal]

Let’s calm down on the whole concern trolling, astroturfing and so on personal accusations. The exact story and motive behind my posting is exactly as was written in that first reply.

It’s somewhat amusing you attack me for failing to research the nitty-gritty of various middlemen shell structures you’re involved in. Unlike you, I’m not paid to do this. I’m only a random crypto user with extensive public social media posting, much of which shows I’m not RocketPool’s biggest fan. Is this the moment where I say “if you had taken the time to look up my history before posting…”?

Or perhaps we could act like adults here, and realise people have different areas of knowledge and expertise.

Mine was a simple question that could have been addressed “based on cold hard facts”, without emotional talk or manipulative language.

Personally, I’m not sure it’s the most convincing rebuttal to have a Lido guy (or whoever you are) join a forum he’s never been on solely to answer an open question mostly directed towards Karpatkey. Talk about aggressive damage control. I came here with a genuine inquiry, I come out thinking you guys are fishy as hell.

[Viszla]

Ill start by mentioning my respect for Karpatkey and what they do. Professional Treasury management in crypto needs to be expanded and is an incredibly important service. I support them and their “tireless work ethic.” That said, I’m troubled that you would say “There is no conflict of interest… because none of their mandates are secret.” That isnt how conflicts work. Their work for Lido is a conflict. It leaves them open to questions regarding their impartiality. Given the danger of a protocol (any protocol, not Lido specifically) controlling >33% of staked ETH, my argument is that an effective Treasury manager would look to reduce the risk that Ethereum faces by reducing Lido’s dominance. We do not know if Karpatkey’s conflict is causing them to underallocate to other LST’s, but both their conflict of interest and your response make it very easy to make the case that they are being influenced by personal economics. As Karpatkey stated, they have sole control of Treasury allocations. Continuing to ignore Lido’s dominance is, in my opinion as a member of GnosisDAO, reckless, and so I am speaking up about it. PhiMarHal, Im sorry you were caught in the crossfire here. Bringing up facts and asking questions should never be punished in a transparent system.

[adcv]

I go one step further actually, the benefit of a transparent system and what makes me so passionate about its importance, is that it’s not only their mandates that are public (as you point out, an insufficient condition), but the entirety of their work output for all of their clients is fully public. There is nothing to disclose as everything is disclosed. The argument of an economic incentive modifying their priorities can be verified. If you took the time to read the structure of the proposal it would be clear there is no such incentive modifying impetus. You can know what you want to know, it is all public and transparent and independently verifiable.

The framing of ‘conflict of interest’ is easy to make when what you require is just the appearance of, rather than actual, impropriety. With similar ease, therefore, I suggest the framing can simply be dismissed. My own responses are immaterial.

[Viszla]

The issue here is less whether they are responding to economic incentives and more that they are failing respond to a credible threat. Lido dominance is a threat to Ethereum. To my mind, there can be absolutely no reason to allow >99% of our Treasury ETH holdings to contribute to that dominance. If Karpatkey is responding to the economic incentives to Lido, that is bad. If they are failing to realize the dangers posed by a dominant LST (>33%) that is equally bad. If you are passionate about a decentralized system, why are you defending its greatest centralizing force? Why are you allowing Ethereum to be one state sponsored move of governance capture away from failing as a system? Why does it seem that so much of the crypto ecosystem has a blind spot when it comes to the ability for Lido to take down this whole experiment in decentralized finance? Lets both be passionate and protect Ethereum.

[Karpatkey]

We appreciate the kind words and constructive feedback in this discussion. Since the very beginning, karpatkey has always committed to acting in our DAOs’ best interests. This involves exposing our execution to the public, listening to the critics, and adjusting according to the fact-based arguments to ensure we keep evolving our practice, i.e. treasury management and the industry as a whole.

Growth is a double-edged sword: it expands your reach and potential to leverage synergies to increase your impact, but it also opens up the way for scrutiny and all sorts of questioning. For this matter—the conflict of interest’s topic naturally surfaces as a consequence of the positive-sum game scenarios in play—something we accept, take seriously, and prefer to deal with using an objective approach rather than emotional rhetoric.

For the record, we acknowledge the importance of decentralisation to ensure Ethereum’s resilience and credible neutrality. We’ve inherited that value from our early days at Gnosis—and will always factor it into our decision-making progress—so long as it doesn’t compromise our mandate’s main goals as GnosisDAO’s treasury managers.

We are also cognisant of Lido’s potential centralisation vectors—and so is Lido—so much that most of the protocol’s tech investment has been on the launch of the Staking Router. This upgrade will move the operator registry to a modular architecture, which will act as a platform for stakers, developers, and node operators to collaborate and advance the future of decentralised staking.

Now, the facts. When assessing portfolio allocation on LSDs, we factor in the following objectives ordered by level of priority:

1. Protect the integrity of the treasury through an extensive risk management practice.

Beyond protocol soundness (of which Lido’s would necessarily come first given its battle-tested infrastructure), risk is mostly a factor of market liquidity. The ability to exit a position with minimum slippage is paramount to us, especially given the magnitude of our order execution. For that matter, let’s consider the scenario offered in this proposal: moving 50% of the current stETH holdings to rETH, which accounts for ~35,000 rETH. Using DefiLlama’s liquidity tool, which sources price quotes on ten different DEX Aggregators at different levels and displays the results over a chart, the slippage for a ~35,000 rETH trade would be ~19%.

For comparison, the slippage for a ~55,000 stETH trade would be 0.24%.

2. Grow the Gnosis Chain ecosystem through finance.

For this matter, we explore ways to leverage the LSD’s network effects created through volume and expanded integration in DeFi protocols to push for the expansion of the Gnosis Chain. This is performed either by fostering the Gnosis Chain traffic through added use cases but also by improving the long-term value of GNO as the native network security’s guarantor. Examples:

• We’ve just launched the wstETH/WETH balancer pool on GC, and are in the process of balancing it and increasing its liquidity; and
• The treasury’s large position in stETH’s Maker vault was a valuable contribution to prove our commitment to the protocol and push for the onboard of GNO as collateral.

We’ve also held positions in Stakewise’s sETH2 in the past to support Gnosis Chain’s only Liquid Staking protocol, and we plan to add new ones as soon as the protocol upgrades to V3 on the Gnosis Chain (which should happen after Shapella):

• Stakewise’s sETH2 deposit 1
• Stakewise’s sETH2 deposit 2
• Stakewise’s sETH2 deposit 3

3. Generate non-income revenues to support the DAO.

When possible, we look for ways to optimise the returns on our LP positions by increasing capital efficiency. This comes mostly down to an active search for the best APRs, incentives, or composability through protocol integrations e.g. LP Boosted Rewards; Leveraged Positions; etc.

Source: https://defillama.com/lsd

From a capital efficiency and yield optimisation perspective, there have been some slight, while significant, differences that make stETH stand out as a better investment so far.

Source: https://dune.com/nesk/lsd-performance

We also monitor the market for arbitrage opportunities, as was the case with Ankr’s aETHc (now ankrETH), for which we were able to capitalise on the market conditions by swapping for ETH in order to optimise yield and diversify the portfolio. This opportunity is creating a ~15% return whether by swapping or eventually claiming the staked ETH.

---

As stated in our previous reply—portfolio allocation at karpatkey is an active process and revised on a weekly basis—so we’ll continue to monitor rETH and other LSD protocols in ways that maximise the objective function described above.

[Viszla]

Thanks Karpatkey for the detailed analysis. If I can summarize, Gnosis DAO’s usage of near 100% stETH comes down to liquidity and yield, with some Biz Dev opportunities to signalling our alignment with the LIDO behemoth. That makes sense to me, but does highlight the issue with ceding complete control of Treasury management to the mandate of economics and growth with no consideration of Ethereum protocol health and safety. This reinforces the need for the DAO to be able to override the profit mission with votes to protect crypto generally.

In short, I agree that we cannot expect Karpatkey to lead the charge to protect Ethereum from regulatory capture and move away from Lido. But I do believe it is reasonable to ask Karpatkey not to move against motions to dictate broad changes to Treasury management when the Treasury management tool in question (stETH, or any LST with +25% market share) is an existential threat to Ethereum.

I’ll end this comment with my favorite New Yorker cartoon. Ethereum is our planet, lets please not fuck this up.