[Constitutional] DVP Quorum for ArbitrumDAO: Implementation & Parameters

[Constitutional] DVP Quorum for ArbitrumDAO: Implementation & Parameters

Abstract

The ArbitrumDAO recently expressed strong interest in moving towards a delegated voting power (DVP)-based quorum model. Consequently, this upgrade, if implemented, will define quorum as:

Quorum = min{max quorum, max{ɑ*DVP, baseline quorum}}

where ɑ, baseline quorum, and max quorum are constants.

This proposal aims to finalize the implementation and accompanying parameters of this upgrade - ɑ, baseline quorum, and maximum quorum. If this temperature check is successful, the proposal will move towards an on-chain vote in February.

Motivation & Context

As it stands, ArbitrumDAO’s quorum is computed based on the total voteable supply, which has no relationship to the tokens registered to vote. This proposal introduces a better approach by basing quorum on delegated voting power (DVP) - a metric that more directly represents the amount of ARB participating in governance and available for voting.

All of the relevant context and motivation may be found in the previously published posts on this topic. The first report focuses on the rationale for the change, the historical performance of quorum, and the security implications of this upgrade. The second report is a comparative analysis of quorum across other large DAOs, as well as shareholder voting in public companies, aimed at presenting insights into ArbitrumDAO’s position relative to comparable systems. The motivation around this update is captured in the initial temperature check.

Specification

Definition

This AIP proposes upgrading to the following quorum computation logic:

Quorum = min{max quorum, max{ɑ*DVP, baseline quorum}}

where ɑ, baseline quorum, and max quorum are constants.

How the Formula Works

The proposed formula works as follows:

  • When ɑ* DVP is lower than baseline quorum, baseline quorum applies.
  • When ɑ* DVP is higher than baseline quorum and lower than max quorum, ɑ* DVP applies.
  • When ɑ* DVP is higher than max quorum, max quorum applies.

In short, baseline quorum and max quorum form fixed lower and upper bounds for quorum.

Proposed Parameters

We suggest the following parameters:

For constitutional proposals:

  • {ɑ = 0.5; baseline quorum = 150m ARB; max quorum = 450m ARB}

For non-constitutional proposals:

  • {ɑ = 0.4; baseline quorum = 100m ARB; max quorum = 300m ARB}

A Trail of Bits audit of the proposed changes can be found here. Note that to initialize the running total, an estimate of the total DVP must be provided as part of the upgrade proposal. A followup proposal can correct any error in the initial DVP estimate, if needed.

Rationale for Parameter Selection

(i) Participation threshold (ɑ)

For non-constitutional proposals, our recommendation of ɑ = 0.4 ensures continuity with the current non-constitutional quorum. At the current DVP of 348.61m ARB, non-constitutional quorum will equate to 139.4m ARB, which is very close to its current value of 145.56m ARB.

For constitutional proposals, our recommendation for an ɑ value of 0.5 aims at creating a safer buffer between voter turnout and quorum while retaining a high voter turnout. Most large DAOs operate with quorum thresholds well below fifty percent of delegated voting power, and corporate and legislative systems rarely employ supermajority quorum requirements. As concluded by research report #2, when quorum is expressed as a percentage of DVP, its value at ArbitrumDAO (~62% of DVP) is roughly double that of the next highest DAO benchmarked. Even after the proposed upgrade, ArbitrumDAO’s quorum would remain higher than all comparable DAOs reviewed by the report.

(ii) Baseline Quorum

Historical participation data of ArbitrumDAO provides strong support for the baseline quorum values proposed. Over the last two years, the average quorum for non-constitutional proposals amounted to ~104m ARB, which motivates our recommendation of a 100m ARB baseline for non-constitutional proposals. For constitutional proposals, the average quorum, excluding periods in which quorum was structurally difficult to reach, falls around 156m ARB, motivating our recommendation of a 150m ARB baseline. These values preserve continuity with observed voting behavior while creating a sufficiently large lower bound for quorum.

Note that baseline quorum values will not apply until DVP drops below a certain threshold. For the constitutional baseline quorum to apply, DVP would need to drop below 300m ARB. At all values of delegation below this threshold, a constitutional proposal will require a quorum of 150m ARB. For the non-constitutional baseline quorum to apply, DVP would need to drop below 250m ARB.

(iii) Maximum Quorum

The implementation also includes an explicit smart contract setting to configure a maximum value of constitutional and non-constitutional quorum. This is a static parameter that can be updated through a DAO vote if needed. A maximum quorum of 450m ARB for constitutional proposals and 300m ARB for non-constitutional proposals are proposed. These values are fully consistent with the current maximum quorum values (where they are defined as 3% and 4.5% of the maximum voteable token supply) and as such aim to retain the status quo on the upper bound. For additional context, DVP would need to go over 750m ARB for the non-constitutional max quorum to be triggered and to over 900m ARB for the constitutional max quorum to be triggered.

Going forward, these thresholds may further act as checkpoints for the DAO to reconsider the DVP-quorum model’s performance in a high-DVP scenario.

ArbitrumDAO Constitution Adjustments

Due to the upgrade’s nature, the following updates will be made to the Definitions and DAO Proposals and Voting Procedures sections of the ArbitrumDAO Constitution, in order to reflect the new way quorum is calculated.

Change 1: Definitions

Old Text

Definitions:

  • AIP: An Arbitrum Improvement Proposal
  • ArbitrumDAO-governed chains: The Arbitrum One and Arbitrum Nova chains and any additional chains authorized by the ArbitrumDAO
  • DAO Treasury: All $ARB tokens held in a governance smart contract governed directly by the ArbitrumDAO and/or the Security Council of The Arbitrum Foundation via on-chain voting mechanisms.
  • Governed Chains: Any ArbitrumDAO-approved chains that are governed by the $ARB token
  • Non-Governed Chains: Any ArbitrumDAO-approved chains that are not governed by the $ARB token
  • Votable Tokens: All $ARB tokens in existence, excluding any tokens held by The Arbitrum Foundation and any unclaimed airdrops

New Text

Definitions:

  • AIP: An Arbitrum Improvement Proposal
  • ArbitrumDAO-governed chains: The Arbitrum One and Arbitrum Nova chains and any additional chains authorized by the ArbitrumDAO
  • DAO Treasury: All $ARB tokens held in a governance smart contract governed directly by the ArbitrumDAO and/or the Security Council of The Arbitrum Foundation via on-chain voting mechanisms.
  • Governed Chains: Any ArbitrumDAO-approved chains that are governed by the $ARB token
  • Non-Governed Chains: Any ArbitrumDAO-approved chains that are not governed by the $ARB token
  • Votable Tokens: All $ARB tokens in existence, excluding any tokens held by The Arbitrum Foundation and any unclaimed airdrops
  • Delegated Votable Tokens: The sum of Votable Tokens currently delegated and eligible for governance participation
Change 2: DAO Proposals and Voting Procedures

Old Text

Phase 3: DAO votes on AIP, on Arbitrum One (14–16 days): During this Phase 3, the ArbitrumDAO will be able to vote directly on-chain on a submitted AIP.

An AIP passes if the following 2 conditions are met:

  1. More Votable Tokens have cast votes “in favor” than have cast votes “against” (“Threshold 1”); and
  2. In the case of a:
    • Constitutional AIP, at least 4.5% of all Votable Tokens have cast votes either “in favor” or “abstain”; or
    • Non-Constitutional AIP, at least 3% of all Votable Tokens have cast votes either “in favor” or “abstain” (collectively, “Threshold 2”).

New Text

Phase 3: DAO votes on AIP, on Arbitrum One (14–16 days): During this Phase 3, the ArbitrumDAO will be able to vote directly on-chain on a submitted AIP.

An AIP passes if the following 2 conditions are met:

  1. More Votable Tokens have cast votes “in favor” than have cast votes “against” (“Threshold 1”); and
  2. In the case of a:
    - Constitutional AIP, at least 50% of all Delegated Votable Tokens, as well as a minimum of 150 million $ARB and a maximum of 450 million $ARB, have cast votes either “in favor” or “abstain”; or
    - Non-Constitutional AIP, at least 40% of all Delegated Votable Tokens, as well as a minimum of 100 million $ARB and a maximum of 300 million $ARB, have cast votes either “in favor” or “abstain” (collectively, “Threshold 2”).

Voting Options & Timeline

This temperature check will include two voting options:

  • Adopt the DVP quorum model, alongside proposed parameters
  • Do not adopt the DVP quorum model, alongside proposed parameters
  • Abstain

We aim to take this proposal to a temperature check vote starting 05/02/2025. Following a successful temperature check, the proposal will be put to an on-chain vote.

8 Likes

Appreciate @Arbitrum for putting together this AIP and the thorough research reports that preceded it. At Tally, we recognize the value of aligning quorum with actual governance participation rather than total voteable supply. A DVP-based model better reflects the reality of who’s available to vote, and the proposed parameters strike a reasonable balance between accessibility and security.

Should this constitutional AIP pass down the line, we’re committed to working closely with the Foundation to ensure the new quorum thresholds are clearly communicated from the governance application. This includes displaying the dynamic quorum calculations so delegates and token holders can easily understand what’s required for proposals to pass under the new model.

Looking forward to seeing this move to temperature check, and subsequently to an on-chain vote.

1 Like

We will be hosting a governance call to discuss this proposal on Wednesday.

DVP Quorum for ArbitrumDAO: Implementation & Parameters (Open Discussion)
Wednesday, February 4 · 15:30 – 16:30
Time zone: UTC
Video call link: https://meet.google.com/djy-endq-gwq

3 Likes

I ran the numbers based on the current top delegates and their delegated voting power.
The goal was to understand how many large delegates would realistically be needed to reach quorum under the proposed DVP quorum parameters.

Using the top20 delegates (total combined voting power ≈ 232.22M ARB - the data may differ slightly due to the constant movement of tokens, but this is not significant), we can understand that this is 95% of all the votes that usually vote.

So, we can use only these vote power to see the following:
Non-constitutional proposals
Quorum = min{max quorum, max{0.4 * DVP, 100M}} = max(0.4 × 232.22M = 92.89M, 100M) = 100M

Constitutional proposals
Quorum = min{max quorum, max{0.5 * DVP, 150M}} = max(0.5 × 232.22M = 116.11M, 150M) = 150M

Cumulative totals show that quorum exceeds 100M after only 6 delegates and exceeds 150M after only 10 delegates

In my view, this creates a very low practical barrier.

One of the few structural mechanisms that gives smaller delegates meaningful influence is the difficulty of reaching quorum. Even if a delegate holds a modest amount of voting power, their participation can matter when quorum is not easily guaranteed.

If quorum can reliably be reached by 5 to 6 large delegates, the incentive for smaller delegates to participate decreases. Once quorum is clearly secured, marginal votes feel less impactful. This may increase voting apathy rather than reduce it.

Reaching quorum should not be trivial. It is supposed to require broad participation. That is part of what makes governance resilient and inclusive.


While I support the DVP model and it is conceptually simpler than the common proposal-based quorum model, the current parameterization may unintentionally centralize power and reduce the incentives for smaller delegates to participate.

Lowering quorum may solve one problem, but it may create another.

I would strongly encourage further discussion on whether the baseline thresholds are set at the right level - I think it should be increased.

5 Likes

@cp0x I agree with the points you’ve raised and alsoshare your concerns. Lowering the quorum to a level where 5 or 6 top delegates could effectively meet it on their own would undeniably diminish the influence of smaller delegates, who represent the majority of the voter base which will discourage participation and weaken the broader governance ecosystem

That said, this is roughly ten months too late since Arbitrum Foundation outlined its new strategic vision which explicitly shifts greater authority and operational control toward Arbitrum Aligned Entitiees (Offchain Labs, Entropy, Arbitrum Foundation, etc.). In that context, the proposed quorum change appears consistent with the broader direction they have already signaled: a more centralized governance structure anchored by those AAE.

2 Likes

I support the proposal and the parameters as well.

I still find great incentives for smaller delegates to have influence, as voting against a proposal that already has reached quorum signals to the rest of the DAO that not everyone is in alignment, and bystanders/token holders might shift their delegation based on this.

1 Like

Interesting approach, clearer execution tooling is definitely an area the DAO can improve. Looking forward to seeing how this develops.

The computation in your post is a bit strange.

Why is it computing quorum based on 232.22m ARB? Quorum is computed as a percentage of the total delegated voting power which is 349m ARB.

This is ~139m (40%) and ~174m ARB (50%). Non-constitutional is quite close to current value of 145m.

100m and 150m represent safety values for Quorum, and for it to reach that number, total delegated voting power will need to be reduced and that’ll likely impact the voting power of most delegates.

I agree that this is consistent with what the Foundation wrote, but they unilaterally proposed it as a direction for DAO, among other things, and DAO didn’t vote for this development.
And I’ve consistently said that centralization is bad, and I’ve always voted against it – because any government strives to do good for itself.
Therefore, DAO’s power is beneficial for both the project and its users.

1 Like

Thanks for the correction - I agree with your calculations.

If we calculate based on your data, it turns out that 139 million ARB would be garnered by 9-10 delegates, and 174 million ARB would be garnered by 13 delegates.

This isn’t significantly different from my initial calculations – we still see that just 10 delegates are enough to pass almost any positive decision, and even fewer are needed to block a decision.

I don’t like the direction voting is taking, where 10 people will decide the fate of everyone else.

voting Against on this offchain vote because parameters should be 0.5 for Non-constitutional proposals, as in, they need majority participation of FOR + ABSTAIN of all delegated voting power to be approved and execute, and 0.67 for Constitutional proposals, as in, they need super majority of participation of FOR + ABSTAIN of all delegated voting power to be approved and executed. As of now, these proposed parameter values are arbitrary and cherry picked to make it easier to pass constitutional proposals breaking the precedent of the current values and current governance hardness in Arbitrum DAO.

1 Like

I am voting FOR this proposal.

It make sense to shift the quorum requirement to delegated supply, as this accommodates investors who wish to hold the token without actively engaging in governance.

While I initially had concerns regarding the ceiling, I now recognize it’s al protection mechanism. It ensures that outsiders with significant capital cannot freeze the DAO, which is essential in such a highly competitive environment.

Here is the recording of the DVP Quorum for ArbitrumDAO: Implementation & Parameters (Open Discussion).

I voted Against

As I mentioned earlier, I’m sure this decision will lead to:

  • apathy in voting (there’s no need to put in the effort if the top 13 delegates provide a quorum)
  • due to apathy - centralization (this is bad, since a small number of delegates control a huge community)
  • instead of incentivizing voting, a simpler method of reducing the quorum is chosen every time (in the eight months since the DVP idea was voted on, no other coefficient options have been proposed)

I am voting in favor of this proposal on Snapshot.

Essentially, it aligns the quorum with the pool of tokens eligible to vote rather than the total supply. I find it highly logical: as delegated voting power increases, the quorum requirement scales accordingly.

I also find the parameters to be reasonable, requiring a 40% DVP threshold for non-constitutional proposals and over 50% for constitutional ones.

Lastly, it is worth noting that this dynamic quorum proposal aligns with the OpCo team’s efforts to expand the number of active stakeholders and significantly increase the total DVP within the DAO. In this regard, I disagree with the following statement:

Under the proposed system, the influx of new delegates with fresh voting power will increase the nominal value required to reach quorum (as the DVP expands). This, in turn, increases the theoretical number of delegates needed to achieve that quorum, hence increasing the relevance of so-called small delegates

2 Likes

Thanks for putting this together. Overall, I think the move to a DVP-based quorum is the right direction and long overdue. Tying quorum to the supply that actually shows up and participates makes far more sense than anchoring it to total voteable supply, which has always felt disconnected from reality.

The formula itself is clean and intuitive. Having both a baseline and a ceiling strikes a good balance between safety and flexibility. It protects the DAO in low-participation periods without letting quorum drift into unreachable territory as delegation grows.

On the parameters, the non-constitutional side feels well calibrated. Using α = 0.4 to preserve continuity with current quorum levels is sensible, and the baseline at 100m ARB lines up nicely with historical participation. This should make the transition feel smooth rather than disruptive.

For constitutional proposals, I agree with the reasoning behind a higher α. A stronger buffer here is justified given the stakes, and even at 0.5, Arbitrum remains stricter than most comparable DAOs. That said, I think it would be helpful to be explicit about how often the 150m baseline is expected to bind in practice and whether the DAO sees this as a long-term floor or something to revisit once delegation patterns mature further.

I also appreciate the inclusion of an explicit max quorum. Keeping the upper bound consistent with today’s effective limits avoids introducing new failure modes, and framing these thresholds as future checkpoints for review is the right mindset.

One small point I’d like clarity on is the initial DVP estimate used at deployment. While a follow-up correction is possible, it would be good to understand how sensitive quorum outcomes are to that starting value and what margin of error is considered acceptable.

Net-net, this proposal feels careful, data-backed, and pragmatic. It improves governance without over-optimizing or reinventing the system, which is exactly what a constitutional change should aim for. I’m supportive of moving this forward and would be interested in revisiting the parameters once we have real data from a few cycles under the new model

2 Likes

I’m voting FOR.

I’ve said this before during the initial discussions, delegated ARB = real votable supply, and that is what our quorum should be based on.

I see the concerns in the comments about the parameters (0.4 and 0.5) and the fears of centralization or “10 people deciding the fate of the DAO.” I hear you. But honestly, apathy is a bigger threat, especially with the market crashing. If we set the bar so high that we can’t pass constitutional upgrades without a lot of effort every single time, we are wasting valuable people’s time.

1 Like

Entropy is supportive of transitioning to the DVP model for quorum calculations. As noted in the Arbitrum Foundation’s research, voteable token supply and therefore the constitutional quorum requirement is continuing to increase at a rapid rate, while the amount of delegated ARB has remained relatively flat over the DAO’s lifetime.

We see some delegates concerned about the reduction in the number of top delegates needed to reach constitutional quorum in the DVP system. With the current quorum calculation method, the 18 largest delegates are required to meet constitutional quorum and top 10 delegates for non-constitutional. While the DVP system will reduce the number of top delegates required to reach constitutional quorum to 13 based on current estimates, this is not an unprecedented level for the DAO and is roughly in line with how the delegate base was composed in its first year.

The continued rate of quorum increase not only risks placing the DAO in a position of not being able to pass important constitutional proposals, but it also places more power in the hands of top delegates. We are approaching a situation where only 2 or 3 of the top delegates can effectively block a proposal by not voting. So while based on the current composition of voting power, the number of top delegates needed to pass a constitutional proposal decreases, the DVP system reduces the boycott risk and therefore arguably makes the system more decentralized and robust than its current state.

Similar to @pedrob, our team also finds the initial parameters chosen by the Foundation to be reasonable. Constitutional quorum should be a threshold that requires broad consensus, but not something that is so challenging that it requires proposals to be essentially unanimous support. Additionally, while the DVP system does introduce different attack vectors or risks, such as apathetic delegator attack, we believe the maximum threshold helps limit this risk. While it is not entirely optimal that another constitutional proposal is required to change the parameters, Entropy reasoned that it is preferable to keep this power in the hands of token holders and governance rather than delegating the authority to the AF or OpCo.

To conclude, we view switching to a DVP quorum system helps align the threshold with actual governance engagement, ensures the system remains functional over time, and preserves the DAO’s ability to govern effectively.

1 Like

We are voting AGAINST the proposal.

The triggers are a way to protect Arbitrum’s governance while also adapting it to a different scenario, namely, a decline in DVP. Setting both minimum and maximum bounds is also a good design choice.

However, in our view, using DVP Quorum is still a way of working around Arbitrum’s difficulty in reaching quorum by reducing the security of its governance.

In the original proposal, quorum was defined simply as 100M/150M (non-constitutional/constitutional). Now, the DVP needs to drop to lower levels (300M/250M) for the quorum to be reduced to those values.

Even under the current configuration, it is extremely “cheap” to attack Arbitrum, given the DAO’s TVL and treasury. The cost to reach quorum has fallen by 67% since this proposal was first discussed, from $56M to $18.3M for constitutional proposals.

Therefore, while the change to the proposal is an improvement, we believe Arbitrum should be more concerned with the DAO’s economic security, not just with making proposals easier to pass.

2 Likes

We will vote “For” this proposal.

We support the shift to a DVP-based quorum model and believe aligning quorum with delegated voting power is a necessary improvement over the current supply-based system. This change better reflects real participation and helps address the growing liveness risk faced by the DAO.

Based on our prior analysis using historical participation data, percentage-based quorum thresholds in the proposed ranges strike a reasonable balance between governance liveness and meaningful participation. In particular, we found that ranges around 42–50% of DVP for non-constitutional proposals and 48–56% of DVP for constitutional proposals are broadly effective, and the Foundation’s proposed parameters are directionally aligned with these findings.

While we view the parameters as an initial configuration that should be revisited as delegation patterns evolve, we believe this proposal represents a constructive and necessary step forward for ArbitrumDAO’s governance.

1 Like