Should the DAO Create COI & Self Voting Policies?

Proposals
1

Edits August 15th

  • Clarified language regarding abstaining to also include neutral voting. ie. voting for all candidates in an election equally to effectively abstain.
  • Clarified language on “Responsible Voting”

Abstract

This proposal aims to restart the conversation around the creation of self-voting and conflict of interests policies. The proposal outlines various options, ranging from maintaining the status quo to enforcing strict self-voting policies. A Snapshot vote to gauge the DAO’s sentiment will be posted on Thursday, August 15th. Please note, this is an actual temperature check, just to gauge the DAO’s opinion on the subject. Based on the results, certain policies may be included in a future delegate of code of conduct as part of an effort to professionalize the DAO’s operations.

Background

Strict policies regarding conflicts of interest (COI) and self-dealing are standard practices in traditional corporate industries. However, with crypto just now emerging from an early stage, entirely avoiding conflict of interests is challenging. Many individuals are either investors, employees, or contributors across multiple companies/protocols on not only Arbitrum, but other L1/L2 chains.

Several other DAOs have established a code of conduct in an attempt to address these issues. Examples range from being opt-in recommendations (Uniswap), simple pledges (Maker, Gitcoin, Aave), and strict no self-dealing policies & expectations enforced by a council and violations resulting in suspension from the DAO’s platforms (Optimism: Code of Conduct Committee, Optimist Expectations).

While Arbitrum has standard rules of engagement for the forums, it currently lacks a delegate code of conduct in addition to an official policy on conflicts of interest. The only mention of the term in the Arbitrum Constitution is in relation to candidate eligibility for the Security Council.

952×127 8.98 KB

Creating a self-voting policy was first raised by @Immutablelawyer at the beginning of 2024 following the LTIPP council elections. While the post resulted in a good discussion, many delegates at the time indicated a strong preference to not move forward with any policy that restricted a delegate’s ability to vote. Thus a temperature check to gather the larger DAO’s sentiment on this issue never occurred.

In recent weeks, the topic has reemerged in conversations, and as part of a series of proposals focusing on professionalizing the Arbitrum DAO’s operations, Entropy Advisors is looking to restart the conversation on a proposed self-voting and COI policy in order to reach a definite consensus. Policies that are voted are favored by the DAO will be included in a future delegate code of conduct proposal.

Rationale

We would argue that within a DAO’s operations, the majority of instances involving COI are the result of an individual/company using their delegated voting power to engage in self-voting or “self-enrichment”. This can exist when a delegate votes for themselves to get elected to a committee, thereby receiving financial gain, or a delegate votes in favor of a proposal that directly benefits themselves or their company. There are certainly other situations where COI can arise such as oversight, but for the purposes of this proposal, the focus is centered on the practice of self-voting.

For clarity and the purposes of this conversation we will lean on the definitions that @Immutablelawyer established back when the topic of self-voting was previously discussed.

  • Conflicts of Interest: A situation where a token-holder or their close-associates stand to directly benefit from the outcome of a proposal.
  • Self-Voting: The practice of where token holders use their voting power to vote on proposals that involve or benefit themselves.
  • Self-Enrichment (Self-Dealing): The act of increasing or enhancing one’s own personal financial wealth as a result of one’s own vote.

As mentioned above, the majority of comments indicated that it was unwise to restrict a delegate’s ability to vote in any manner. To paraphrase the discussion, hesitation to implement a self-voting policy boiled down to primarily two reasons:

  • Philosophical Objections: Delegates are entrusted by token holders to best represent their own values and interests. Forcing a delegate to abstain from a vote due to COI concerns prevents the delegate from truly representing their base’s interests and thus disenfranchises the token holders delegating.
  • Lack of Enforceability: Even if self-voting was monitored manually by contributors or a committee, it is quite easy to circumvent self-voting policies onchain by simply transferring voting power to anonymous wallets. To quote @DisruptionJoe, in general, “unenforceable rules create situations where bad actors have an advantage over good actors’'.

While both are strong objections, there are positives to implementing COI and self-voting policies. First, they help establish the “rules of the game”, so that every participant in Arbitrum governance is operating under the same set of expectations. Furthermore, disclosure requirements can help lead to greater transparency while also incentivizing participants to act in an ethical manner. Most importantly, with such policies standard practice in other industries, Arbitrum DAO can improve its image as a professional organization by holding its actors to similar standards.

Specifications

Below are a few policy options the DAO can adopt to help address and mitigate COI and self-voting. The options are presented from least restrictive (do nothing) to what we view as the most restrictive.

Keep the Status Quo: Arbitrum DAO can continue to operate as it has been. A delegate code of conduct does not need to include specific policies on disclosure or self-voting.

Disclosure and Transparency Policy: If a potential COI exists, it is expected that proposal authors or delegates disclose the nature and extent of the interest. While it may not always be clear if an individual stands to gain “directly” or “indirectly”, it is recommended to lean on the side of over-communication in the name of transparency. Delegates are not expected to abstain or alter their voting preferences in any way, but are free to choose to abstain.

“Responsible Voting” Policy in Elections: In an election with multiple seats, candidates may vote for themselves as long as they also cast votes to fill the remaining positions. Put more simply, a candidate should not cast a vote solely for themselves in an election. In the case of a weighted voting election, the candidate’s weight for themself should not be larger than the weights for other candidate. To the best of their ability they should strive to make the weights as equal as possible among the number of candidates necessary to fill each position.

  • Example: I am running in the GCP council election, which has 3 open seats. I can vote for my own candidacy, but must also cast 2 additional votes for other candidates.

This is a policy that has worked well in other DAOs and helps address the natural incentive to withhold votes for competing candidates in an election while also making sure token-holders are not disenfranchised.

Credit to @JoJo for the term “Responsible Voting”.

Strict Self-Voting Policy: In the event of a conflict of interest, under the simple definitions stated in the Rationale above, it is expected that a delegate or proposal author abstains during the voting process. In the event that “Abstain” is not presented as a voting option, it is expected that the delegate votes in a neutral manner as to not impact the results of the vote. For example, in an election a delegate may even divide their voting power in equal weight to effectively abstain. This policy would include candidates being prevented from self-voting in elections and delegates from voting on proposals that directly benefit oneself, company, or interests.

Enforcement

To date, the primary deterrent to egregious breaches of an adopted code of conduct is the potential for reputational damage. Other DAOs have experimented with various enforcement mechanisms in regards to upholding a COI or self-voting policy. One of the best examples is Optimism DAO, which created a delegate code of conduct in 2021 with enforcement carried out by its foundation. In its Season 5 (September, 2023), Optimism created a Code of Conduct Council that took over responsibility for overseeing and enforcing breaches. However, recently this council has been scaled back and its Token House self-dealing policies have returned to being enforced “via the free market for delegation.”

This journey highlights the difficulty of enforcing any type of self-voting policies and the likely ineffectiveness of simply creating another committee to provide oversight. To our knowledge, no major DAO has taken the steps to actually bar an individual from voting at the contract level (not that this is recommended). Enforcement is an area in which Entropy is actively seeking input and encouraging delegates to share opinions and ideas.

Currently, our team is leaning towards making the future delegate code of conduct opt-in, following in Uniswap DAO’s footsteps by creating a document that serves simply as a form of guidance. We envision soft-enforcement can be applied through the Delegate Incentive Program by requiring participating delegates to abide by the ratified code of conduct in order to be eligible for rewards. Further UI changes on Tally could also be applied to help token holders discover which delegates have publicly committed to following the agreed upon code of conduct.

Voting Choices

The temperature check on Snapshot will use approval voting. Each delegate can select any number of choices, each selected choice will receive equal voting power. So for example, if a delegate wishes to see the DAO adopt both a disclosure policy and responsible voting, then they should select both options on the Snapshot vote.

Since this is a proposal that changes the DAO’s current status quo, the following choices have been worded in a manner that each “FOR” vote indicates desire to adopt a certain policy.

FOR: Disclosure Policy
FOR: Responsible Voting Policy
FOR: Strict Self-Voting Policy
AGAINST: Maintain Status Quo
ABSTAIN

Timeline

A Snapshot poll will be posted on August 15th to gauge the DAO’s overall sentiment. This will allow for ample conversation and input from delegates beforehand.

This proposal is part of a series of temperature checks to further improve and professionalize the DAO’s operations. Others include the Improving Predictability and Adopting Shielded Voting proposals. Depending on the outcome of this vote, COI and self dealing policies will be included in a delegate code of conduct. As mentioned previously, a future, more comprehensive proposal will look to combine all of the DAO’s new processes into a trial period, with the successful ideas eventually implemented into the Constitution.

Cost

This proposal carries no cost to the DAO.

5 Likes
2

Thank you for this proposal. Would like to add some nuances to this

The “abstention” should likely be reframed as “neutral voting”. We live in a world in which always voting is +ve, either for current programs (delegate incentive program) or potential future one (staking). In some cases, there is just no way to abstain in a vote in an explicit way, especially in elections.
Neutral voting can be, for example, to vote for all parties (including yourself) in an equal weight. And, bear in mind, might not be unfortunately always possible

This can likely get a clarification as well: the definition does not encompasses all cases. What if we have a weighted voting, and I vote 98% for myself, 1% for candidate B and 1% for candidate C? I am technically complying while achieving the same of just voting for myself.
Rephrasing this as “Put more simply, a candidate should not cast a vote solely for themselves in an election. In case of weighted voting, the weight of voting for themself should not be bigger than the weights of votes for the other candidates” could be a solution. Note that especially this point might be prone to error due to how weights can be expressed in snapshot, and we can expect in future candidate to express voting as per the above with, potentially, allocating 0.x% more to themself just due to miscalculation. As long as this difference doesn’t materially change the end result of a vote it doesn’t likely need to be addressed, especially in this framework in which nothing is enforceable.

Yes: if there is a vote for a protocol to receive a grant or not, and i am part of that team or an advisor I should disclose it in my voting rationale if i am a delegate voting; if I am the author of the proposal, even more. And it’s also natural that if I hold 10% of circulating supply, I should disclose it as well. Likely, if I hold 5% of the supply makes sense to disclose.
Where do we draw the line here? Is it a % of coins? We go in the situation in which it blurs.

We already discussed this point; but I like idea of “it is recommended to lean on the side of over-communication in the name of transparency”, I think some folks already applied this from time to time, but putting it on (virtual) paper helps.

Great job @Entropy

1 Like
4

Hi! Thank you once again for proposing the type of discussions that the DAO needs.

The first thing this discussion makes me ask is, what do we want to avoid? And secondly, what are the undesirable side effects of prohibiting certain conduct?

What we want to avoid:

Voting for oneself per se is not bad. If a delegate is convinced that their proposal benefits the DAO and can demonstrate this with arguments, there is a net positive benefit for the DAO if that delegate votes for themselves or the proposal they are pushing.

Why would someone be forced to vote in a way they believe is less beneficial for the DAO or Arbitrum users?

So, the problem itself is not the direction of the vote but the potential conflict of interest that may arise as a consequence of the vote.

So, despite the explanation…

Although self-voting can sometimes be the most obvious form of conflict of interest, it is not always the most manifest in practice. The transparency of votes makes it easy to see who is operating solely for their own benefit without adding any value to the DAO, which ultimately would be the most objectionable action.

What are the undesirable side effects of prohibiting certain conduct?

Suppose we establish the premise that self-voting = COI and therefore prohibit it. In that case, we will be forcing delegates to vote for options that are not optimal or necessarily the best by default.

According to the Arbitrum Constitution: “The Arbitrum ecosystem should be managed for the benefit of all Arbitrum users.”

And that has to be the guiding principle that determines the direction of the votes. There must be a net positive outcome from the proposals for all Arbitrum users.

Another undesirable side effect would be the one already discussed in the post by immutable lawyer, and raised by @DisruptionJoe and myself. We would be creating situations with low transparency.

For example, I don’t vote for myself for a position in the DAO, but I coordinate with another delegate with similar voting power to vote for me in exchange for my vote for them. What would that be? These situations are encouraged by such a prohibition.

For this reason, I am against a rule that directs the votes of the delegates. Even though these rules may seem noble and well-intentioned, in practice, they make the DAO less transparent.

Also, I think this is not necessarily the biggest problem. Conflict of interest situations manifest during the discussion and voting stages of a proposal, but they materialize and are only possible during the execution of the proposals.

What do I mean by this?

Consider a proposal where voting is taking place to onboard a service provider to perform a specific task (e.g., @Entropy).

Is there a COI if Entropy or its members vote in favor of the proposal? There are arguments to support both yes and no (the latter being the viewpoint I agree with more).

Now, during the execution of the proposal, it will become clear whether a COI actually existed. Essentially, this is because they might fail to fulfill the agreement, take the ARB tokens, and show that they were voting in their own interest at the expense of Arbitrum users.

When it comes to more complex proposals that involve committees, the execution of the decisions made by these committees will reveal whether their members are making decisions based on the greatest benefit to Arbitrum users or using their position to benefit themselves over the users.

How do I think COI should be minimized?

In the first place, I support your Disclosure and Transparency Policy.

But second, I think the DAO should have a consensus on the Minimum Viable Proposal (MVP). A framework should be established outlining the minimum content that a proposal must contain in order to be submitted to Snapshot.

We should work on turning this “Incomplete Guide to Submitting a Proposal” by @Sinkas (which is very good) into a complete guide that includes the necessary requirements for a proposal to be voted on.

Within that minimum content, there should be a section on Conflict of Interest explaining who cannot be involved or take actions related to the proposal due to a conflict of interest.

This is not an enforceable rule that prevents proposals from going to Snapshot (similar to the rule that proposals must be published on Thursdays). However, a significant number of delegates agree on the minimum content a proposal must have to be approved:

  • The work will be more efficient and effective, saving delegates time.
  • Proposals can be rejected solely for not meeting this rule.

By including conflict of interest rules within the proposals themselves, we can determine prior to the vote who cannot benefit from it or be part of it. This way, we are not limiting how to vote but rather what can be voted on.

I hope this makes sense

5

Thanks for sharing the Entropy.

Responsible voting:

Knowing what would/wouldn’t be counted as a Close Associate would be useful. We have seen the greatest friction on this topic around elections where it is unclear which individual candidates a delegate can/can’t support.

For example, would the following be treated as Close Associates for Treasure as a delegate:

  • ARC members
  • Treasure Angel Investor
  • Treasure Community member with significant native governance token holdings
  • Ex-employee of Treasure who stays in regular contact with the team

Also noting my support for @JoJo’s observation on the need for clarification around the specific language for Responsible Voting.

Strict Self-Voting Policy:

This term may exclude many delegates from voting on certain important proposals, for example, future user incentive programs.

For that reason, I believe the Strict Self-Voting Policy is infeasible.

6

We heavily agree with the sentiment of the DAO taking on some sort of COI policy moving forward. Although, we are interested in all of the options proposed, a simple disclosure policy and more responsible voting policies seem to be the proper next step for the DAO. We agree with JoJo in that you should be able to vote for yourself, but maybe the vote should not be larger than the distribution of votes to other candidates. Bare minimum though, we think a disclosure policy going forward would help the DAO become more transparent and foster better communications between delegates.

As other delegates have noted, we would like to see more strict definitions that draw lines in the sand for what defines a close-associate and otherwise.

7

Notifications says that something got linked. I am not paid for posting. Great.

My philosophical objection is that governace might be irrelevant pretty soon. I hope your bags are well pumped. Volatility is high enough to nullify your gains via taxes.

8

TL;DR

COI’s being culturally acceptable makes us look unprofessional and undermines the perceived integrity of our DAO’s governance system to the outside world. This change is long overdue, and can be enforced in some extent by the Delegate Incentive Program.

YES PLEASE!

All of the above.

This would be a massive improvement in the DAO, and will make us look soooooo much more professional. Delegates have an advantage over non-delegates just by being in the room talking with other voters. Abstaining helps make things a bit more even for proposals that don’t have delegate benefit.

Said more clearly: If you are a delegate that has a proposal up for vote, and you abstain, I would argue you still have an unfair advantage to a nondelegate that has a proposal up for vote.

The relationships delegates have with each other, especially large delegates, creates an unfair advantage. Not to mention the fear that voting against someone’s proposal may make them not favor you in some future business/DAO opportunity.

Enforcement

Strong agree that we don’t need another council, this can be mostly socially enforced, however, in practice we do have a council of sorts already in place that could manage some enforcement: The Delegate Incentives Program.

If someone votes in favor of their own direct benefit, do they really need the Delegate Incentive Program’s reward? Seems like they have incentive enough. If we already have a team reviewing all these comments, it should be easy enough for that team to see if people have clear conflicts get called out for not declaring them or not voting abstain… they should get their incentives slashed, if not completely, then by some large %.

Feasibility of Strict Self-Voting Policy

I disagree 100%.

Large generalized incentive programs will have a large burden to overcome, it is true. I personally see this as a good thing. It would mean to pass an incentive program, it would have a better chance if it were more narrowly scoped. This is a bug not a feature. We just passed a vote to stop proposing any incentives for 3 months after these massive STIP, LTIPP and STIP-Bridge proposals. It is very questionable if they were good for the DAO, but there is no question, they were good for many large delegates (especially STIP and STIP Bridge).

IMO Proposal’s that benefit a large number of high ranking delegates should be put to a higher scrutiny. Our current lack of COI policies actually make proposals that benefit large delegates EASIER to pass. Imagine what this looks like to an outsider, “If I just structure my proposal so that the large delegates benefit, it will be sure to pass.”

This is DAOing it wrong, I’m sorry.

5 Likes
9

Thank you @JoJo for identifying a few additional edge cases. We have updated the language of the proposal to hopefully make things more clear in regards to neutral voting and equal distribution of voting power in a weighted election. Additionally, our team is in the midst of another proposal based on other points raised in @Bob-Rossi’s previous RFC to standardize non-security elections. In it we will suggest some other guidelines for proposal authors such as making sure there is always a method to abstain or remain “neutral”.

@pedrob, thank you as well for the thoughtful response. Incorporating COI disclosure requirements into the actual proposal template is an interesting idea and if the DAO wishes, is something we can work to include in the larger DAO operations proposal & delegate code of conduct that will seek to update the Constitution.

@Pepperoni_Jo3 you bring up an excellent point regarding the specific language used to describe COI and close associates. This highlights the nuanced nature of COI discussions and is why we started with a temperature check rather than a full code of conduct policy. The DAO may indicate that it does not want a strict-self voting policy and therefore discussing exact language is not even needed. Overall, we acknowledge that the definitions presented are likely far from perfect, but our hope is that this vote will help indicate the general direction the DAO wishes to take regarding COI.

To reiterate the purpose of this proposal, no policy, guideline, or social contract will be adopted as a direct result of this vote. In addition to sparking a conversation, it is designed to help us understand overall sentiment to inform what clauses and language to include in a future, more comprehensive delegate code of conduct proposal.

4 Likes
10

Incentive program is not needed in my opinion. I have spent a lot of time going through the forum and other materials. I do not think that just pressing a button is enough to get 5000k per month. Get a job.

Effectively, you are getting paid to post a couple of sentences in a completely meaningless way. Closing your eyes on conflicts of interests. Looking for another tranche and training session.

And the incentive program just helps you create a corporate America once again. Fat bellies are fed. Even Usain Bolt would be jealous.

11

I am voting For.
At the Aave DAO we already have this kind of policy and with Arbitrum being such a big DAO there is also the need.
Also the DAO is for example only paying delegates that dont receive any other kind of “fixed” payments for example because they are a risk service provider.
The DAO is probably overspending because there aren’t rules like this to keep it fair and only pay those that are active and don’t receive anything else from the DAO.

2 Likes
12

I wrote about the same thing in the discussions in January. Corporations have documents and a regulator that can track this, but it is quite difficult to do there too.
There is always the option to delegate one vote to your friend, so I think there is no point in limiting anyone.
Of course, I want everyone to be honest, but in this case, honest participants will lose.

Also, I think that this can be easily circumvented by temporarily transferring votes.
Secondly, if a delegate has few votes and his own votes are not enough to be elected, why should he give anything to those who have more of these votes? There is no logic in this and no honesty or nobility.

I think you can’t compare voting in crypto and in real life: in real life, each person is one vote, but here everyone can buy tokens/votes. You can talk about that buying votes is generally unethical, but talking about it in crypto is complete nonsense)

2 Likes
13

I voted “FOR: Responsible Voting Policy” on snapshot because it strikes a balance between ethical governance and practical delegation. This approach helps maintain trust within the DAO while allowing delegates to represent their communities effectively.

Here’s an idea: should we explore tools to automate conflict of interest disclosures, making it easier for delegates to manage and disclose their potential conflicts transparently?

1 Like
14

I’m voting for “Disclosure and Transparency Policy”. Encouraging disclosure can push for openness and trust within the DAO. The outcome is that voters can make informed decisions while maintaining the freedom to vote as they choose. I think that in this way, both accountability and autonomy are guaranteed.

1 Like
15

For all the reasons mentioned above, I am voting for “FOR: Disclosure Policy” and “FOR: Responsible Voting Policy”.

I think is definitely good if people disclose their conflicts; is also good if, in a vote that can benefit themself, they also vote in a way that is “responsible”, so for other 2 seats in a 3 seats election etcetera.
I am against the self strict voting policies, especially because protocols for example should be allowed to vote for proposals that directly involve themself. If that was not allowed any user delegating their precious arb to the protocol of their choice would see their delegation being useless, because the protocol can’t vote for his own benefit. A byproduct that I don’t want, among others.

Regardless of my vote, @Entropy I am also interested about the specific implementation of this proposal. Knowing it can’t be enforced, should any new vote from now on decide what happens if there is a self voting (with the strict voting policy on) or anything that is against the expected behaviour? How we will specifically handle these conflicts?

1 Like
16

I agree with the “FOR: Responsible Voting Policy” vote as it balances ethical governance with practical delegation needs. Maintaining trust while allowing effective representation is crucial. Exploring tools to automate conflict of interest disclosures is an excellent idea. It would streamline transparency, making it easier for delegates to manage and disclose potential conflicts, thus enhancing overall governance integrity.

17

100% agreed with “FOR: Strict Self-Voting Policy”.

This is something we’ve been discussing in forums since the DAO was first formed. Delegates (and protocols) should abstain from proposals that directly benefit them, and disclose the ones that indirectly may. The definition of “indirectly” here should be treated carefully, since it’s only natural that (for example, incentives) have an indirect benefit in the whole ecosystem, actors, and protocols (which, with incentives, is the idea anyway).

18

Blockworks Research will be voting in favor of the Disclosure Policy, Responsible Voting policy, and the Strict Self-Voting policy.

These seem to be the most straightforward solutions currently to conflict of interest and the least disruptive. We can glean incidents/possible violations from these policies and decide if there should be any type of further ruling enforced or guidelines from there.

From our point of view, it’s the Self-Voting policy is attempting to draw a line for where DAO delegates/protocols/users represent their own interests and have the DAO have a distinct set of interests delineated from those. This makes sense, because while individuals in the DAO come from a list of different backgrounds, at the end of the day the DAO should still have a few select distinct interests separate from those individuals. Ideally we can achieve this through incentive alignment, however there are sometimes areas where policy is necessary.

19

Actually, I don’t think the delegate should be allowed to vote on his own proposal, especially when the proposal involves a large amount of fund and the delegate has a lot of votes, because I can’t believe that the delegate is completely selfless. So I think Strict Self-Voting Policy is necessary.

20

I am voting FOR “Disclosure and Transparency Policy”:

I think it makes sense that disclosure of COI is part of delegates’ mandate.

21

I voted for: Strict Self-Voting Policy

I love to see this proposal. I think it was necessary to set some rules and not just leave it at the status quo. The best policy is a Strict Self-Voting policy because it means delegates with bigger power can’t decide or swing votes for their benefit. Other policies are fine but could lead to more bureaucracy and complex processes. Let’s keep it simple and just self-regulate each other.