[Discussion] Safety Module - Governance Impact

[raho]

Authors: Traver, Ryan - Governance Analyst | Messari Governor
Date: Feb. 28, 2023

Simple Summary

This post aims to initiate a discussion surrounding the impact of the Aave Staking Module on the Aave DAO’s governance.

Background

@bgdlabs

It is important to not forget that the SM has a core role in the Aave governance, and the numbers provided seem to support that, with an important percentage of the current AAVE rewards coming back to the ecosystem in one way or another.

We suggest analyzing from the governance point of view too because one of AAVE’s main utilities is its governance power.

The recent discussion surrounding the redesign of Aave’s Safety Module began with a comment from @John_TV_Locke, followed by a discussion created by @Llamaxyz. In response to the latter post, @bgdlabs raised awareness of the fact that any change to the Safety Module would result in a direct impact on Aave governance, where stkAAVE plays a core role.

Introduction

With recent discussions surrounding Aave’s Safety Module underway, @tnorm and I have found inspiration in the feedback provided by BGD. We strongly agree with the comments surrounding the need for proper analysis and due diligence from a governance angle before considering any potential redesign. We decided it would be best to analyze the potential impacts of this implementation in an independent forum post in an effort to not distract from the valuable discussion taking place in regards to Safety Module design.

The goal of this post is to initiate productive discourse surrounding the role of the Safety Module in the Aave DAO’s governance. We aim to achieve this by providing a brief analysis of the current landscape, possible alternatives, and areas to explore further. This post, by design, does not draw a definitive conclusion and is intended to act as a starting point for further community discussion.

---

Summary: The Safety Module’s Role in Aave Governance

The Aave DAO is governed by a combination of AAVE and stkAAVE holders. Aave Stakers are incentivized to remain mission-aligned with the protocol (by risking slashing in the case of a shortfall event) and are rewarded with protocol voting power. By design, this loop incentivizes both protocol insurance and quality governance decisions.

Analyzing Stakers’ Activity in Aave Governance

In its current design, AAVE emissions are distributed exclusively through the Aave staking module on Ethereum Mainnet. Therefore, all AAVE distributed through the staking module is immediately available to govern the protocol. Currently, the circulating supply of stkAAVE is 3,289,288.48, representing 23% of circulating AAVE (14M) (Source). Given that ~20% of AAVE is locked in the Safety Module, it is important to understand the impact that stkAAVE has on the DAO’s governance.

Current Delegation Stats:

• stkAAVE Delegated Voting Power: is 858,230.51.

• AAVE Delegated Voting Power: 571,790.71.

• stkAAVE Delegated Proposal Power: is 1,587,067.68.

• AAVE Delegated Proposal Power: 659,094.54

(Source: Dune - @aave_companies)

Current Governance Participation (On-Chain)

The chart below displays the average of all on-chain proposals’ participation rates for a given month. The participation rate is defined as the sum of voting power that voted divided by the total amount of potential voting power. The highest percentage of voter turnout the DAO has seen was 20.7%, on AIP 14 (Source), with the highest average turnout hitting 5.1%. There are external factors that may play a role in the low turnout, such as the gas fees required to vote. However, it is worth noting that this is significantly lower than other on-chain DAOs, like ENS, Gitcoin, Uniswap, and Compound (who sit above 25% participation), and is even lower than less active DAOs like dYdX (5.3%).

(Source: Messari Governor - Aave DAO Profile)

(Note: The ‘potential votable supply’ (AKA circulating supply) used is likely much less as there are many tokens in LPs, Centralized Exchanges, and Aave’s Ecosystem Reserve. Emphasizing again, a more thorough analysis is needed here.)

Potential Alternatives

The recent discussions on the Safety Module consider whether or not the SM is an effective insurance fund for depositors, not if it is an effective way to reward mission-aligned agents. With that, the initial questions we should be asking are:

1. Are SM emissions optimized for rewarding mission-aligned agents? Or simply for paying for insurance?
2. If emissions are optimized sufficiently, how does that alignment change if the asset deposited into the safety module changes from AAVE to a more liquid asset?
3. What dependencies does Aave’s governance module have regarding the Safety Module? (I.e., what implications will a Safety Module re-design have on Aave’s governance, from a technical perspective?)

As discussed above, incentivizing Safety Module stakers to take on the slashing risk makes sense, and we will leave that mechanism debate to the other thread. However, if changes are made to the SM, it is important for the community to begin thinking of ways to minimize the impact on DAO’s governance.

Some potential ideas that may be worth exploring:

• Independent Staking Module without Safety Module.
• Liquidity Rewards for Aave V3/GHO.
• Bonds for GHO PoL
• GHO safety module
  • stkAAVE rewards could assist with GHO depositor retention via underlying GHO incentives for stkAAVE holders (discounted minting).

---

Further Research: Areas To Expand With Deep Research

Granular Aave Voter Study

While touched on above, it would be valuable to understand the DAO’s voter composition on a granular level to understand the role stkAAVE plays in governance decisions.

Governance Token Distribution

Is there an alternative method to the Safety Module that could entice voters, and perhaps have a greater RoI in terms of governance involvement? It goes without question; the Aave DAO is one of the best run, and higher signal DAOs in the space. However, it may be valuable to at least obtain a deeper understanding on alternative methods for emissions and the possible RoI they may bring in terms of governance.

Alignment of Users and Governance Participants

Are users of the protocol (depositors and borrowers) participating in governance? Do they hold AAVE tokens to be able to participate if a vote would impact their usage?

[eboado]

Good work on the analysis @raho and the Messari team.

@raho

This is significantly lower than other on-chain DAOs, like ENS, Gitcoin, Uniswap, and Compound (who sit above 25% participation), and is even lower than less active DAOs like DyDx (5.3%).

• In practice, 14M is not the total votable supply of Aave at the moment, as there are different venues from where AAVE can’t participate (aAAVE, L2s, AAVE on stkABPT). So I would try to be careful with these assessments, as they lead to really wrong conclusions.
  Combining on-chain and Snapshot, Aave has one of the most (if not the most) active governances in DeFi.

@raho

there no longer appears to be a strong alignment between those staking AAVE and those participating actively in AAVE governance.

• Are you sure the data of stkAAVE participating in governance (as a % of voting power voting on proposals) is correct as under 5% ? I have important doubts about it, and that argument is key for your conclusion

@raho

there no longer appears to be a strong alignment between those staking AAVE and those participating actively in AAVE governance.

• Do you have data regarding the AAVE used to vote (including delegation always) coming from the one distributed as a reward on the Safety Module? And obviously only considering those stakes that decide to keep the rewards.

[raho]

Hello, and thank you for the thoughtful response @eboado, these are great points!

@eboado

In practice, 14M is not the total votable supply of Aave at the moment, as there are different venues from where AAVE can’t participate (aAAVE, L2s, AAVE on stkABPT). So I would try to be careful with these assessments, as they lead to really wrong conclusions.

I completely agree that 14M is not the true votable supply; as we pointed out above, the circulating supply would need to be thoroughly broken down to achieve a true understanding of the % of votable supply in which stkAAVE makes. See:

@raho

(Note: The ‘potential votable supply’ used is likely much less as there are a large number of tokens in LPs, Centralized Exchanges, and Aave’s Ecosystem Reserve. Emphasizing again, a more thorough analysis is needed here.)

Additionally, I’d like to emphasize that we intentionally did not come to a conclusion. Our goal was not to propose a change but to initiate a discussion around the topic. With this in mind, we suggested further research and analysis on the topic instead of a conclusion.

---

@eboado

Combining on-chain and Snapshot, Aave has one of the most (if not the most) active governances in DeFi.

For context, Aave’s Snapshot strategy includes:

• aAAVE
• AAVE + stkAAVE
• Matic AAVE + amAAVE
• AAVE in stkBPT

We did not combine Aave Snapshot with Aave on-chain because the multiple strategies that are relevant to Aave’s snapshot space lack relevance in on-chain voting. It would make sense that more voting strategies would increase participation, but that does not necessarily mean that executable proposals will benefit from this or that Safety Module stakers are participating any more in governance than before. The rewards should be optimized with on-chain participation in mind, as that is what drives actionable protocol decisions.

---

@eboado

Are you sure the data of stkAAVE participating in governance (as a % of voting power voting on proposals) is correct as under 5% ? I have important doubts about it, and that argument is key for your conclusion

The stkAAVE used for this was taken from delegated stkAAVE events, which does not account for stakers voting with personal holdings. We recognize that the latter is an important aspect of the analysis and would need to be included in further voter analysis. However, we wanted to post this discussion as a starting point to gather ideas and pinpoint the critical areas for analysis if further research is needed. See:

@raho

Granular Aave Voter Study

While touched on above, it would be valuable to understand the DAO’s voter composition on a granular level in order to clearly understand the role stkAAVE is playing in governance decisions.

This point could be refined and deeply analyzed in a deeper governance analysis, which we recommend before concluding it as a concrete fact (again, emphasizing the open-ended discussion format and need for extended research).

---

@eboado

Do you have data regarding the AAVE used to vote (including delegation always) coming from the one distributed as a reward on the Safety Module? And obviously only considering those stakes that decide to keep the rewards.

Regarding the AAVE rewards used to vote, it still appears minimal. A rough overview of the reward claimers’ activity in terms of governance is pictured below.

(Source)

This graphic shows a small number of stakers claiming rewards and voting (or delegating), but does not give us the total amount of votes sourced from the SM. This data is available, however, it would also need to be refined to display the exact AAVE that was used to vote (sourced from SM rewards), as a % of the ‘votable supply’ at the time of each vote.

[eboado]

@raho

The stkAAVE used for this was taken from delegated stkAAVE events, which does not account for stakers voting with personal holdings. We recognize that the latter is an important aspect of the analysis and would need to be included in further voter analysis. However, we wanted to post this discussion as a starting point to gather ideas and pinpoint the critical areas for analysis if further research is needed.

Yes, but the data is simply wrong.
In terms of absolute amount (and % of voting power), the delegatees voting on the last proposals have way more stkAAVE delegated than AAVE.
Meaning that the biggest influence on voting power comes from stkAAVE delegators.
So having a The majority of Safety Module stakers are neither active delegators nor have voted in Aave governance is wrong and really misleading.