[Constitutional] AIP: DVP Quorum

Proposals
1

Constitutional

Abstract

This AIP proposes to upgrade the on-chain voting system for the ArbitrumDAO’s governance contracts. The upgrade focuses on changing how the quorum threshold is computed such that it is a fixed percentage of the total Delegated Voting Power (DVP), impacting the quorum threshold for constitutional and non-constitutional proposals.

Motivation & Rationale

The ArbitrumDAO has been struggling to achieve quorum on constitutional proposals. The gap between participation and quorum has been shrinking and has now become alarmingly low.
If no tokens are excluded from voting, quorum currently increases at 36M ARB/year for non-constitutional proposals, and 54M ARB/year for constitutional proposals. It will soon reach a point where active voting power available is no longer greater than quorum, bringing the DAO to a standstill. This proposal aims to avoid such a scenario and provide a long term fix to ArbitrumDAO’s quorum issues.

The situation at hand is fueled by two fundamental issues. Firstly, quorum is computed based on the total voteable supply which has no relationship to the tokens registered to vote. Secondly, delegated voting power has remained relatively unchanged over time. Together, this has led to a scenario where the quorum threshold is increasing, but the total voting power is steady. This proposal focuses on addressing the first issue by redefining the quorum computation logic. Increasing DVP is a separate endeavour that should be pursued to further strengthen Arbitrum governance.

Kindly refer to this recently published research report and Dune dashboard that explores the current situation in greater detail. The report supports moving to a DVP-based quorum model as it has several advantages including alignment with voter registration and not requiring token holders to actively exclude their votes from the votable supply. A DVP-based solution also preserves the fundamental properties of the existing voting system such as one token one vote and the ability for an external party to easily evaluate whether quorum has been reached.

Specifications

Proposed Quorum Model

This proposal seeks to replace the current voteable-supply-based quorum definition with a Delegated Voting Power (DVP)-based definition:

Quorum = max (ɑ*DVP, baseline quorum), where ɑ and baseline quorum are constants
  • ɑ will have different values for non-constitutional and constitutional proposals
  • Similarly, baseline quorum will be different for non-constitutional and constitutional proposals

The formula ensures that quorum varies as a fixed percentage of DVP (ɑ*DVP), unless the baseline quorum is greater. The graph below shows how quorum would change with DVP if we choose an ɑ value of 0.5 and a baseline quorum of 150M ARB.

1
11580×980 84.2 KB

The research report recommends choosing an ɑ value in the [0.4, 0.5] range with non-constitutional and constitutional quorum tending towards the lower and upper end of that range respectively. The report suggests 100M ARB as the baseline quorum value for non-constitutional proposals and 150M as the baseline quorum value for constitutional proposals. However, we will refrain from asking the DAO to vote on the exact quorum parameters at this stage. Parameters will be confirmed through a separate Snapshot closer to the on-chain proposal so that the community has more time to inform its decision.

To implement the proposed solution, the following smart-contract level changes need to be made:

  1. The $ARB token contract needs to be upgraded to keep a running total that is updated on each delegation change and token transfer.
  2. The DAO governor contracts need to be upgraded to use the new total DVP metric. Existing calculations need to be replaced with DVP-based thresholds.

Offchain Labs will be taking on the development and testing of the proposed upgrades. An untested, unaudited PoC of the token contract upgrade can be found here.

Timeline & Voting Options

We intend to take this proposal to Snapshot next week (09 October - 16 October). Through this temperature check, we aim to collect community sentiment in the overall direction. The proposal will include the following voting options:

  • For DVP Quorum
  • Against DVP Quorum
  • Abstain

Should For DVP Quorum be the majority preference, OCL will start working on the implementation, including production ready code and necessary audits. The exact parameters of the DVP Quorum model (ɑ values, baseline quorum values) will be determined through a separate Snapshot closer to the on-chain proposal.

2 Likes
2

I was probably the first person to suggest this approach for Arbitrum DAO quorum problems roughly 1 year ago in the chats…

but…

i feel that… to actually propose to update the token and governor contracts, in this way, without conducting any external governance security audits / war rooms / red teaming / etc, and when no other onchain DAO contract in the world (that actually secures as many billions as the Arbitrum Core governor does) has ever opted for this approach, like ever…

yeah… this is kind of a wild thing to propose like this…

I’ll be voting Against.

And it would be funny if this proposal wouldn’t hit the constitutional onchain quorum required.

3

when no other onchain DAO contract in the world (that actually secures as many billions as the Arbitrum Core governor does) has ever opted for this approach, like ever…

Just because others have not implemented a solution does not mean that no one should. If this was the approach taken in crypto, then we as a community would not have Ethereum, Smart Contracts, Rollups, etc, which were all very controversial and even demonised when they were proposed. Our job is to identify what can work, whether it works, if it is desirable to do, and then try to do it.

in this way, without conducting any external governance security audits / war rooms / red teaming / etc, and

I do not see why a war room is even appropriate here? We are very much taking the security of the system seriously which is why we worked on the report, the proof of concept, will have the final code audited, and continuously reviewed by all of us to make sure there is confidence if/when it is deployed.

And it would be funny if this proposal wouldn’t hit the constitutional onchain quorum required.

The DAO is not here for your entertainment. Please be more respectful* to everyone here. As you can imagine, people do not appreciate remarks like this.

5 Likes
4

Hi, thanks @Arbitrum for moving this proposal forward. We were wondering why the community isn’t being asked to consider the α value at this stage. If this proposal passes, there will still need to be another vote and due diligence on the percentage later. If that follow-up were not to pass, this proposal might have limited effect. Would it perhaps be more effective to present some percentage ranges upfront, so the DAO can evaluate both the direction and parameters together? We’d be keen to understand the reasoning here.

5 Likes
6

Besides Curia’s comment about no one proposing Alpha options, I’d like to understand the following:

  1. Who can change these parameter (Alpha) and how? Is a vote required?
  2. What threshold values ​​can be applied?
  3. What should be done if the negative forecast about high DVP and the impossibility of achieving a quorum with a certain Alpha value suddenly materializes? Is there a procedure to lower this Alpha? (Such a situation is easy to imagine if, for example, AAVE adds delegation capability to their contract)
8

Thank you for the question, @curia. We need alignment on the new direction before the exact parameters can be voted on. Coming to the DAO with a fixed set of parameters and asking it to vote might have felt like we were expecting a rubber stamp on the new quorum model - which is not our intention.

A temperature check on the broad DVP-Quorum model has two benefits: first, it gives the DAO more opportunities to discuss the model and potential parameters; and second, it allows the community to validate this research direction, meaning the implementation work can start in parallel. You can find the recommended value range in our research report to further understand the nature of the proposed change. We will incorporate feedback from this temperature check and the following discussion into our final recommendation and vote.

2 Likes
9

I’m sure you’re not putting these example innovations in the same pedestal as this DVP Quorum proposal. Because they are not comparable, and even more importantly, they are moved by very different motivations.

It seems to me, more and more, that the motivation behind this proposal is to make it even easier to pass constitutional proposals. To the point where we would be able to pass existential protocol changes with very little legitimacy to do so. When all ARB is unlocked in a few years, we would be able to change the protocol with just a 150M ARB quorum, which would mean that just 1.5% of the tokens would be making a decision that impacts all token holders. That’s a very, very small percentage, for constitutional changes in Arbitrum.

I’m not talking about contract code security. I’m talking about increasing the surface for governance attacks. Just in the last call, I highlighted one potential new governance attack that this quorum method enables for less than half the price of what it would currently cost (according to @amanwithwings). And it’s the kind of attack that our direct competitors would be able to pull off pretty easily and with total deniability.
Obviously these are sensitive things to discuss in public, but since Arbitrum would be the first DAO in the world, that secures as many resources as we do, to adopt this quorum method, I think it is very much appropriate to run a war room with industry leaders in this regard.

Maybe even setup a public bounty for governance security researchers to report new vulnerabilities enabled by this quorum method.

I meant that it would be funny as in, it would be ironic.

And maybe it would force the AF to finally address the root of the issue, which is not how the quorum is calculated, but the fact that the largest token holders don’t care to delegate their tokens because there’s no incentive to do so.

Also, we did pay @CastleCapital and @Nethermind to research this issues a few months ago.
And their conclusion was:

And their recommendations were:

and

And that’s why I’m running my own kind of incentives for ARB delegation to me, in the open:

2 Likes
13

The parameters will be fully controlled by the DAO. Any change will require a constitutional vote.

Can you kindly clarify what you mean by this question?

If we consider the “Apathetic Delegator Attack” identified in the report and what appears to be what you are asking about, i.e., significant delegation, but the voting power is not active, in order to raise quorum and prevent any proposal from passing.

There are only two solutions to it:

  • More ‘active’ delegations that can outpace the ‘inactive’ delegation.
  • Security council is requested to tweak the configuration.

Generally speaking, delegation is never unintentional, and any design that promotes the growth of DVP should also include mechanisms to ensure active participation of delegated tokens.

Yes, the motivation behind this proposal is to lower the quorum for constitutional proposals as currently there is a mismatch. Your arguments above assume that DVP (which quorum will depend on) and turnout will remain the same. If DVP is 300M ARB, then quorum will be 1.5% of total supply, if it is 1B ARB, then quorum will be 5%. It remains to be seen where DVP stabilizes once token unlocks are complete.

We do not understand what you mean by sensitive things to discuss in public. Security of a system should be evaluated by as many people as possible which is why we also run a public bug bounty program, pay for audits for any contract changes, and work on reports (like the forum post, or academic papers) to outline how a system works alongside potential security issues that may arise.

In the call, what you described is essentially the apathetic-delegator attack which is summarised towards the end of our initial research report. The cost of attack calculation that yields a value lower than the current model assumes the typical day-to-day level of apathy witnessed in the DAO today. If an attack did surface and became known to us all, we suspect the apathetic profile will change among token holders.

15

The recording of the DVP-Quorum for ArbitrumDAO: Open Discussion can be found here.

16

OK, thank you for finally saying it plainly. And by “mismatch” you mean mismatch between what exactly? Between what the Arbitrum DAO constitution currently says and what some stakeholders would hope for? Mismatch between what?

Well of course, but Delegated Voting Power (DVP) has been going sideways for 2 years now, right?

My main point, and the point of others as well, is basically:

Why don’t we try to get DVP to 1B ARB first, and only then we will see if we need to change how the quorum is calculated?

Because this argument of yours is basically saying that for constitutional proposals to keep the same quorum hardness they originally had, the 5% of total supply, we would need to triplicate the amount of delegated tokens under this new DVP style quorum being proposed, while the amount of delegated tokens and turnout have remained roughly the same throughout the whole history of Arbitrum DAO. This is, I’m sorry to say it like this, a ludicrous expectation.

17

We have voted AGAINST this proposal

Changing how quorum is measured is a delicate matter.

Below, we describe why, before any change is considered, attempts should be made to improve voting power delegation/delegated token actively participating on governance on Arbitrum.

Decrease in Arbitrum Security

As the core of blockful lies in our security-oriented approach, we’ll begin by bringing a governance security perspective to the discussion:

Currently, the quorum for proposals is 3% and 5% for non-constitutional and constitutional proposals, respectively. Looking at the Votable Supply, this equates to 141M and 212M $ARB. In terms of value, this means that to reach the quorum for a proposal in the Arbitrum Treasury, it costs $56M; in Arbitrum Core, $84M.

Considering the liquid assets (specifically in “ETH Holdings” - 11.021 ETH Held on Treasury U$49M and/or Non-native DAO Allocated assets U$183.34) in the Arbitrum treasury, and over $1B in $ARB — both stored in the Arbitrum Treasury — the cost to reach quorum is low. Mainly because the number of votes obtained in Arbitrum proposals does not exceed 240M votes in 2025. In many cases, it did not even reach 200M $ARB.

With the current proposal, considering ~50% of Delegate Voting Power as the quorum for constitutional/non-constitutional proposals, we set the quorum at 100 million/150 million $ARB.

We know that ɑ will be discussed in the future. But if we set ɑ to a value below the current suggestion, Arbitrum makes it even cheaper to reach the quorum for a proposal in the DAO, by ~30%.

If it costs $40M to reach quorum on Arbitrum and the DAO has more than $50M in liquid assets - disregarding $ARB - we will have a profitable attack on the DAO.

Just a reminder:

Attack profitability is not merely a mechanism design vulnerability — it’s a business model.
Here’s an example from ENS, where we identified and mitigated an attack that could have drained $150M in liquid treasury with an initial investment of only $3M.

[Temp Check] Governance Security: Compensating blockful for preventing a potential attack on the ENS DAO - Temp Check - ENS DAO Governance Forum

Of course, the DAO has protections against this—the Security Council, for example - and reaching quorum does not mean approving a proposal.

But we want to make it clear how Arbitrum reduces the security of its governance, simply because it has no tokens in circulation being delegated—a problem that needs to be solved first.

Risk of Repeating the Current Problem, Considering DVP

We understand that the idea of considering Delegate Voting Power for quorum is to have a metric that varies according to delegations. Thus, with less delegated power, we will have a smaller quorum and, theoretically, less difficulty in approving proposals in the DAO.

The problem is that with DVP, there is a risk of the same problem occurring: paralysis/standstill in the DAO.

Since DVP is considered for the quorum, it is expected that all delegates with voting power will participate. Especially the majority, who are responsible for helping proposals reach their quorum. However, those who help approve a proposal may be the same ones who can block it.

If the majority delegates (top 10) do not vote, and DVP is considered for the quorum, there will be an even greater difficulty in reaching the amount necessary for the proposal to be approved.

In numbers: the 10 largest delegates represent 128M $ARB, ~40% of the Delegated Voting Power. If they fail to vote, and 50% of the DVP is used for quorum, it would be almost impossible to coordinate all other delegates to reach the necessary 150M $ARB. This is because there would only be 189M $ARB delegates left to vote.

It is an extreme situation, but again, such a delicate change in governance needs to consider the extremes. That is why the Security Council was created: to protect the DAO and Arbitrum from attacks.

To try to solve this problem without changing how quorum is measured, one alternative is to encourage $ARB holders to delegate their tokens to active delegates in the DAO.

With more DVP, and less delegates participating, we will have a quorum with an inflated number - uncorrelated with the reality.

Encouraging Delegators, Not Just Delegates

There is currently a movement to encourage inactive governance token holders to delegate their tokens to delegates with a high participation rate in a DAO.

  • blockful is designing an incentive program for the ENS DAO, where delegators are rewarded with governance tokens for delegating their tokens. The profitability of delegation depends on several factors, such as the delegate’s recurring participation.

  • Event Horizon proposed a similar system for Uniswap, but restricted to a certain number of delegates with a high degree of participation in the DAO.

Regardless of the model, both focus on one problem: the increase in circulating supply, the stagnation/decrease in delegation, and the decrease in DAO participation.

While Arbitrum tried to take steps in this direction — such as the (Re)delegation Week — the effective and sustained impact of these initiatives has yet to be achieved. There are more efficient mechanisms that, with slightly greater coordination and effort, could deliver far stronger and longer-lasting results.

What is (Re)delegation Week? (+Application thread)

With more tokens in circulation, mechanisms must be created to attract delegations from people who choose to invest in tokens and can earn an annualized return simply by keeping them delegated to an active DAO participant.

Can this be seen as a cost to the DAO? Yes, but it is the price of increasing the cost of improving the DAO’s economic security.

This model can be discussed in Arbitrum governance, and all delegates can verify whether this is a good idea for the DAO. We see it as an alternative to try to improve the relationship between DVP x Votable Supply/Circulating Supply without necessarily changing the way the quorum is measured.

We are available to discuss this topic with the community and also to include Arbitrum on our governance risk assessment dashboard.

1 Like
18

yeah, the current cost to reach quorum is already low, and it just became even lower with this recent drop in $ARB price

1 Like
19

gm, voted FOR.

I think a more dynamic model makes sense and I trust the research done on the matter, with the ability for the DAO to discuss the parameters before the final decision.

As Amen explained in the call today, the quorum is not the principal factor for a governance attack - it’s just a minimum condition. What matters is how much honest voting power there is.

20

Thank you for the proposal!

This measure seems positive as it ensures smoother DAO operations and addresses a structural issue, namely the growing gap between total votable supply and delegated voting power, which threatens governance paralysis.

I will vote FOR, as the rationale is sound. Still, we must make sure decentralization isn’t compromised.

According to @Arbitrum :

Continuing the previous line of thought, my personal position is that not only the exact parameters (e.g., the α factor) should be discussed, but also any additional ideas should be co-designed with open delegate participation. In such a serious change, every willing delegate - if there are any - should be accepted in proposing their ideas, and if these ideas are valid, @Arbitrum should remain open to adopting them. There’s no reason to rush such an important change. Both the final proposal and the process itself, in such an important proposal, should reflect the DAO’s democratic principles.

21

The following reflects the views of L2BEAT’s governance team, composed of @krst and @Manugotsuka, and is based on their combined research, fact-checking, and ideation.

We voted AGAINST.

While the current quorum model needs improvement, we are not convinced that this proposal is the right solution for the identified issues.

First, it does not address the underlying issue of tokenholders not delegating their tokens to participate in governance. I recall that this issue was raised during a DAO meeting in Denver in February of this year, where the conclusion was that there would be an initiative to activate approximately 300M ARB for active voting power. Shortly after, there was indeed a small spike in delegated ARB, but it was nowhere near the discussed values, and its effect has since diminished. Therefore, we would first like to better understand whether these initiatives to activate more ARB voting power have failed or simply were never implemented.

Next, the proposed solution represents a significant change that would require upgrading two of the most important governance contracts. This process would be inherently costly - it requires not only smart contract development, but also thorough audits and updates to all tooling that currently relies on existing calculations, such as Tally, Proposals.app, and various dashboards displaying these values. We recall a similar situation in Optimism almost two years ago, where governance contract upgrades resulted in numerous governance tools breaking across the ecosystem.

We are not suggesting that upgrades should never be made, but they should not be undertaken lightly. If we proceed with one, we should consider what other features could be included in such an upgrade. From the top of our minds, examples include partial delegation and private voting or flexible voting—both of which were identified as missing features necessary for the proper implementation of staking mechanisms.

At the very least, a proper impact analysis - along with a detailed plan for addressing potential issues and a thorough cost breakdown - should be presented before requesting approval for this approach.

Furthermore, we share some concerns about potential additional security issues and governance attack vectors that have been raised previously in this discussion.

Finally, we believe that alternative mechanisms should be explored before adopting a complete redesign. For example, the Auto-Abstaining Wallet model, implemented by Scroll DAO, presents an interesting approach. It mitigates quorum issues without requiring fundamental contract upgrades, thereby limiting the impact on external systems and reducing additional security risks. Arbitrum could benefit from a more thorough analysis of this model before committing to an approach as impactful as DVP Quorum.

We support continued research and open discussion, but do not believe this proposal, in its current form, should move forward at this time.

2 Likes
22

I’m voting in favor of this proposal. Active governance is essential for the DAO’s progress to continue, and this change to base quorum on DVP addresses is useful for avoiding potential deadlocks.

That said, the comments from @krst are worth noting, they highlight the need for more efforts to activate undelegated tokens and raise valid concerns about implementation costs and disruptions. However, I don’t think these detract from the proposal’s benefits. A proof-of-concept is already in place, and the Arbitrum team has the expertise to handle the upgrades efficiently without major issues.

2 Likes
23

I will be voting “For” as I believe that having a dynamic quorum that is based off actual delegated voting power makes more sense given the DAO environment. It’s clear from now years of DAO governance (across all DAOs, not just ARB) that it’s basically impossible to get new token holders to actively delegate their tokens. My fear is that we are going to get to a point where we can’t meet quorum and will be faced with a choice of gridlocking the DAO or making a one-time exception - neither of which will be great.

I do believe continued effort into making delegating easier and more popular should be explored, but with the DAO now being a few years old and all attempts at this so far being a failure… voting against an adjusted quorum in hopes of a better solution coming along is more wishful thinking then an expected outcome.

1 Like
24

We are all aware that Arbitrum has a quorum problem.

In fact, as referenced by @paulofonseca, this was researched by ourselves and @Nethermind and presented to the DAO in these two reports:

We understand the desire to tie quorum to delegated voting power rather than the circulating supply of ARB and the benefits this brings. However, there are several concerns we have around the implications of such a change, which we believe need to be appropriately assessed.

Firstly, it does not recognise the growing issue of voter apathy or seek to address the stagnation in the DAO’s delegated voting power - something we know there is a clear need for and should focus on before discussing altering quorum calculations. We have seen informal delegation events happen in the past, and encourage the DAO, the Foundation and OCL to do more in this regard.

Secondly, the proposed solution is entirely novel and untested, and therefore likely has not had various edge cases hypothesised and tested for attack surfaces. In addition to various ‘edge cases’, we expect models to be created and simulations run on simple price action and tokenholder changes. If we were to experience a major black swan event or a prolonged bear market, which led to a mass exit of ARB tokenholders, each quorum calculation would impact the DAO in vastly different ways.

Thirdly, market research on other available and tested methods should be fully assessed, analysed, and critiqued in the context of Arbitrum to fully inform both the Foundation and delegates of their benefits and risks as valid alternatives. For example, @krst references:

Lastly, we will just highlight the valid point @krst made about the breadth of the undertaking to upgrade key contracts, with the potential to introduce new vulnerabilities, but also that this may be an opportunity to discuss what else should be upgraded more widely.

Whilst we understand the motivation behind the proposal, we encourage a broader research and discussion framework that can lead to a unified approach to solving quorum, voter apathy, and delegation issues once and for all. In light of this, the DAO should be able to fully understand the benefits, risks, and alternatives to such an implementation.

As such, we will be voting AGAINST this proposal.

2 Likes
25

voting Against on this offchain vote because we need to research the governance security implications of this change way more seriously.

26

I meant what limits will be set in the contract to avoid accidentally setting 0 or 1000%.

This is simply a technical question to protect against inappropriate thresholds.