DAO Discussion: Vote Buying Services

[Arbitrum]

Last weekend, hitmonlee.eth paid 5 ETH (~$10k) on LobbyFi for 19.3M ARB votes (~$6.5m) and all votes were cast for CupOJoseph.

The purchase to ‘vote for option 8’ can be found here:

Arbitrum Transaction Hash: 0x851a568966... | Arbitrum One

LobbyFi has been active in the ArbitrumDAO for several months, but this is the first material example of someone willing to pay a significant amount (i.e., 5 ETH) to influence the outcome of an election.

There are other examples, such as ~20.1m votes purchased for 0.0652 ETH to vote in the previous TMC Recommendation proposal, alongside several occasions when the voting power was less significant at ~1m ARB and purchased for less than $20.

Purchasing votes in the OAT elections should be viewed as a precedent-setting moment for the ArbitrumDAO, and it warrants a discussion on whether (and when) vote-buying-as-a-service platforms should be allowed.

We’ll take this opportunity to walk through LobbyFi’s logic, why it is a precedent setting moment, potential defenses available to the DAO, and next steps in regards to the OAT election.

Of course, our entire discussion should be independent of CupOJoseph who also expressed concerns with how LobbyFi set their pricing for this election.

Finally, we decided to wait until the OAT election was over to kickstart this discussion, to avoid any adverse effect it may have had on CupOJoseph’s participation.

LobbyFI’s Logic For Enabling Vote Buying on OAT Election

LobbyFi publishes the motivation for how the price is set for every proposal including the OpCo election:

"Since the candidates have been pre-approved by the Foundation (see post Update on OAT Elections’ Candidate Eligibility Evaluation phase), we regard this proposal as secure enough to go live on LobbyFi. We will therefore make the voting power available for 1 (one) candidate only, without voting power splitting. To encourage the community to amplify their choice in this important election, we will make the auction available, where, following the winner-takes-all principle, the pool that will gather the highest amount of ETH will get 100% of the VP. Also, we must stress that the bids that are made for an option that will not turn out to be the highest can be withdrawn.

The potential upside for a candidate is the term’s compensation ($7.5k * 12 months ~ 47.1 ETH) and the bonus at current prices (100k ARB ~ 18.7 ETH) → ~ 66 ETH per candidate. LobbyFi will charge a flat 5 ETH for an instant buy, meaning that the auction may be won with 0.5 ETH and upwards, by the largest pool; otherwise, a vote equal to abstain will be cast."

The main reason that LobbyFi enabled vote buying for the OAT election appears to be due to the pre-approval process run by the Foundation. It should be highlighted that the compliance process run by the Foundation checked candidates against the following logic:

• Applicants/nominees must be individuals
• Required to disclose and maintain a public list of all actual and potential conflicts of interest, including but not limited to financial, personal, DAO governance, and professional
• May not be a direct representative or full-time employee at network competitors (e.g., Solana, Polygon, Optimism)
• Required to pass a KYC process and enter into an Agreement with the OpCo legal entity. The Agreement will cover, including but not limited to, appointment, conflict of interest, mandate, confidentiality, record-keeping, duty of impartiality, recusal, self-dealing prohibition, and ethical trading standards.

Unfortunately, the process is based on loose criteria on whether a candidate may have a conflict of interest, but it doesn’t necessarily place judgement on whether the individual will make a good candidate for the OAT as that decision was reserved for the ArbitrumDAO via the voting system.

Precedent Setting Moment

The fundamental issue in an election is that a potential candidate may get elected not because the active community of delegates voted for them, but because someone purchased enough votes to help swing it in their favour.

This raises an issue of legitimacy for the OAT council as their job is to represent the DAO in the running of the OpCo alongside holding accountable the execution of future proposals in the DAO.

Interestingly, on the delegate chat, @EventHorizon decided not to turn on the voting system for this election, with the following comments:

• “We elected to stay away from elections as avoid contention of outcomes and support smooth/clean an election process as possible."
• “It is just ideologically better to support clean democracy process with as few points of content as possible. (NOT stirring the pot, not your style, I understand).”

However, the real question that needs to be asked, is the following:

• Should the ArbitrumDAO permit people to explicitly purchase, and compete to purchase, votes?

We could shy away from the elephant in the room, but unfortunately, it does set a precedent that vote buying is permissible. If the DAO does not discuss it, then the silence encourages it to continue.

If we evaluate the long-term trajectory of vote-buying-as-a-service, the incentive is to onboard as much voting power as possible, and to sell it to the highest bidder. There is a reality where a vote-buying service has enough votes to pass or significantly swing in their favour any proposal in the DAO.

On the flipside, as we can see in the current case with the large token holder who was willing to delegate ~20m votes to LobbyFi, vote buying does provide a way to activate otherwise inactive voting power. Additionally, like in the OAT elections, it can transparently reveal whether votes were purchased.

Above all else, in DAOs, the goal should be to implement a governance process that can lead to positive outcomes for the community. We should discuss whether selling the right to make decisions to the highest bidder aids the decision making process and ultimately benefits Arbitrum.

DAO’s ability to disqualify votes

One way for the DAO to defend itself against vote buying is to disqualify votes if it is discovered that someone has purchased votes from another party.

If the DAO wishes to enforce the disqualification of votes:

• Ignore votes from the final tally on Snapshot,
• Require DAO proposals to send funds to a trusted multisig (like the MSS) which can be returned if votes are disqualified.

Votes for technical upgrades cannot be disqualified without intervention by the Security Council to step in and veto the proposal during the 8 day time lock window. If we want other mechanisms to disqualify votes, then the governance protocol needs to evolve to include a checks and balances system that can veto decisions made by token-weighted voting.

Next Steps

We (The Arbitrum Foundation) are not convinced that vote buying leads to better decision making for the DAO. In fact, it can be argued that vote buying undermines the very purpose of DAO governance which is focused on making a decision amongst a set of informed and engaged delegates. If it was up to us, we’d discourage its use and update the code of conduct with new rules related to vote buying and enforce penalties whenever indisputable evidence became available.

However, this is not a decision for the Arbitrum Foundation to make. It is up to the Arbitrum DAO, its collection of contributors, delegates, and stakeholders, to discuss the issue at hand and come to a solution that works for you.

We invite all community members to discuss it with a particular focus on:

• The concept of vote-buying-as-a-service, both in relation to the OAT elections and to any proposal in the DAO, to reach a conclusion on whether it should be permissible and under what circumstances, if any.

We hope there is a substantial discussion around this topic. If necessary, we will follow up with a temperature check vote to help ratify the outcome and any potential decisions that may need to be made.

Finally, it is never a boring day in the Arbitrum DAO and this is another big moment for us. To our knowledge, no other DAO has had an explicit vote-buying system of this nature or size before. We should treat the past few months as a learning exercise and use it to help inform what should happen in the future in regard to vote buying.

[tamara]

We (the ARDC) are currently researching this very topic and expect a first draft of the research conducted by @DefiLlama_Research end of next week.

Based on the engagement of this thread we could also look at facilitating a DAO discussion around it.

[paulofonseca]

@Arbitrum

We should treat the past few months as a learning exercise and use it to help inform what should happen in the future in regard to vote buying.

We should indeed.

I believe that proceeding with this course of action will just turn this into an endless cat and mouse race.
Because vote-buying, in delegated token-weighted governance is inevitable.
Trying to censor vote-buying by deny-listing some votes, will only force these platforms into doing it in a much more obfuscated way, that will be way more difficult to track.

And at least right now, when a vote is bought on LobbyFi, that information is instantly public. In fact, I believe that that was the only reason that empowered the delegates in the DAO to then respond to a vote being bought, by strategically voting on other candidates like many shared in OpCo – Oversight and Transparency Committee (OAT) Elections

So in my point of view, what we should learn from the past months exercise is that we should look into embracing vote buying as part of our governance process.

LobbyFi is being open and transparent right here in the forum, as to how they are pricing their votes. it would be great to see the @Arbitrum Foundation, @offchainlabs and all other delegates engaging with their posts (like this one for the OAT) so we can get better decisions on how to adequately price each vote.

[cupojoseph]

@Arbitrum

Of course, our entire discussion should be independent of CupOJoseph who also expressed concerns with how LobbyFi set their pricing for this election

Thank you for making this distinction and waiting until after the election to post I’m very grateful for this and I think it speaks to high degree of integrity found within the AF.

To add some more context to the discussion:

1. I did not ask for these votes
2. I (and everyone else) had no way of verifying the votes for OAT were even cast for me, because the election was shielded (private) on snapshot.
3. I voluntarily disclosed that Hitmonlee told me the votes were cast for me. Actually, I think my choice of transparency greatly negatively effected my candidacy. But I’m content with the results anyway.

---

In regards to actually addressing this, I have posted in the delegate channel and on twitter publicly (the following is my message):

Telegram

Other mitigation strategies that have been proposed (by me):

• increase DAO voting resilience by increasing quorum requirements, or remove abstain from quorum Basically, play with this more and update it more often as the DAO evolves BEFORE this becomes an issue, not after.
• Support major positive contributors to the DAO so they dont feel like vote buying is the only way to get their work through
• provide other yield to ARB token holders so they arent 100% incentivized to use vote buying.

We can address these concerns and protect the DAO long term without banning lobbyfi or other similar mechanisms. Goodharts Law… I actually don’t think you could ban vote buying so long as ARB is a tradable, transferable token with economic value, and by trying to ban it we will just drive these actions underground into the dark forest. Dark Pools, private zk payments, etc.. Vitalik wrote about futarchy more than 10 years ago and vote buying has been something considered by society for at least 500 years.

This is the reality of all DAOs from now on, so lets seriously address the root cause of the issues. To those reading this, If you want to address the underlying issues and DAO resilience, please consider my above proposals.

[paulofonseca]

@cupojoseph

I have posted in the delegate channel publicly:

just to clarify that the Arbitrum Delegate Telegram Chat is not public. messages in that chat should not be shared publicly without the consent of the authors of those messages.

on the other hand, LobbyFi has a public telegram channel and there has been quite a bit of conversation in there, exactly about this topic, that can be seen at Telegram: View @lobbyfichat

[0xDonPepe]

Censorship isn’t the solution.

LobbyFi just exposed something that already happens all the time in DAOs: lobbying. The only difference is that now it’s transparent and traceable.

Prohibiting vote buying won’t stop influence, it’ll just push it further into the shadows, where it’s even harder to audit. Behind closed doors, people lobby each other in private DMs, Telegram chats, calls. With LobbyFi, at least we know who’s buying what, how much they paid, and for what purpose.

This isn’t corruption; it’s transparent lobbying with a paper trail.

At the end of the day, let’s not kid ourselves, this DAO is a plutocracy. We literally buy a governance token to have a say on proposals. That’s the entire premise. LobbyFi just builds on that logic by making influence tradeable, visible, and direct. If you can buy ARB to vote, why not rent it? Why is one acceptable and the other taboo?

If anything, platforms like LobbyFi could counterbalance whales by enabling smaller actors to collectively amplify their voice. And if a whale wants to lease out their power, why not? It’s better than it sitting idle or being influenced off-chain.

We don’t need knee-jerk bans. If we go down this road, where do we draw the line? Will we start policing who can buy tokens before a snapshot? Scrutinize every large transfer? One could argue it’s actually more legitimate when someone spends their own money to gain voting power, unlike treasury-held tokens that were never purchased (airdropped to VCs or the Foundation) and are now being delegated behind closed doors, with zero transparency around who gets them or what conflicts of interest might exist.

[GFXlabs]

Our thoughts are still evolving on this topic, but we do want to point out two things:

1. This cannot be prevented in practice, so it’s not clear that censure of vote-buying out in the open, and with an audit trail is correct. There is very little to stop someone from privately giving $10k to an ARB holder or a delegate, and in that case governance would not even be aware of it.

2. These votes were purposefully put up for sale by the ARB tokenholder. That makes this harder to classify as an abuse compared to borrowing the votes (there are 10m available to borrow on Aave alone) where the ARB holder may not be aware their votes can be rented. One can also imagine a delegate changing a vote for compensation, and none of that compensation flowing through to the actual tokenholder. At least in the case of LobbyFi, it is the tokenholder knowingly renting out the votes and also receiving the benefit rather than a protocol or delegate.

Perhaps this deserves a separate conversation, but it also, the AF has become increasingly opinionated in DAO politics, and we do not believe statements like the one below are in keeping with the role of moderator and neutral facilitator. The referee shouldn’t have a favorite team.

@Arbitrum

If it was up to us, we’d discourage its use and update the code of conduct with new rules related to vote buying and enforce penalties whenever indisputable evidence became available.

If the AF wants to propose policy or otherwise act in a role that is politically active and not strictly ministerial, we would recommend spinning out or creating a separate role similar to the governance facilitators that exist at MakerDAO.

[olimpio]

In open, permissionless systems, whoever can create an incentive or have an edge will do it, and it should not be censored. If attempted to be censored, new ways will be created, and it will become a cat a mouse game. It’s not like this is entirely new - governance tokens can always be borrowed before a proposal, delegated, and used to vote. And lobbying already exists in the private setting.

Isn’t banning or establishing if VP purchasing is enabled or not is analogous of ethereum censoring MEV? Or to HL coordinating with specific validators to liquidate an exploiter at a certain price point different than what the oracle is reporting? Centralized decisions to patch inefficiencies are not the way: incentives and mechanisms put in place by the DAO should push desired behaviour. In this particular case for example, if there was another alternative for idle ARB (take a loan against it, generate yield, use it in DEFI, stake it for some reward) it would not be delegated to a VP marketplace.

Lobbyfi has got 19.2M ARB delegated to them. 18.9M (or 98%) comes from a single wallet: 0xed40cde6689389b9d4ab7dd11054c57f1050d55a. Is that an investor? He seems to be selling 790K tranches every 30 days:

Wallet’s 39M original ARB tokens were received in Nov 2023. And, in the meantime, it’s is generating yield with Lobbyfi’s kickback. A question for @lobbyfi, is the wallet getting any yield for this delegation? It does not seem to have anything available to claim in the UI.

@Arbitrum

One way for the DAO to defend itself against vote buying is to disqualify votes if it is discovered that someone has purchased votes from another party.

The way this is framed makes it look like the option available characterization for vote buying is that it’s “wrong”, since the DAO needs to “defend” itself from it. The FND’s opinion is clear. From a practical point of view I believe excluding such votes would be very challenging.

This said, I also think it’s valid to subject anything to a vote by the DAO (for example excluding these votes from certain proposals) and although I might not agree with the approach I would most certainly agree with the concensus.

Very interesting developments

[EzR3aL]

This is it, the pricing is the key element imho.
We cannot ban or censor wallets, because LoobyFi could simply use another one. So that doesn’t make sense.
What is important is to engage with LobbyFi and try to define prices for votes that seem important like OAT. Where you definitely want and need someone highly skilled rather than someone rich just buying enough votes, although it didn’t work this time.
What if we as a DAO could vote upfront in a LoobyFi thread for a price range? Because if the DAO agreed to the price, then its a fair mechanism.
That way we are in charge setting the right price range for important votes like OAT, ADRC, GCP and so on.
Not an offchain vote but a simple vote within the forum.

[Hawheik]

As has been pointed out already, I think banning the practice is difficult if not impossible since the same vote-trading actions can still be taken, in the shadows. Perhaps not as economically efficient, but it just means that power is further concentrated among a few well-connected individuals at the detriment of the general public. If vote-trading is happening, I would rather see it ‘democratically’ available to all, and for the practice to be publicly observable to all participants.

That being said, I think LobbyFi’s services provide one unfair advantage we normal users do not have, buying votes from LobbyFi (or by bribing any other user) is the only way for us to acquire more voting power for an on-going vote, as it is happening.

Due to the way snapshots work, each of us only have the amount of voting power available for a given vote that we possessed when the vote was initialized, at the time of the snapshot. None of us can go out and buy more ARB in the middle of a vote to acquire more..except by renting from LobbyFi (or similar service). I think this is an important power imbalance created here.

If we allow LobbyFi or similar services, should we not also allow users to add to their voting power through the other means the market provides? The service LobbyFi provides is roughly analogous to a temporary restricted loan of ARB, but LobbyFi is able to provide a feature-as-a-service (essentially through a technical loophole) that the voting system has fundamentally been designed to deny its users.

Whatever stance we take, in my mind we ought to be fully aligned in both aspects. We should either:

1. Allow LobbyFi to do what they currently do, while also allowing every voter to freely accumulate more ARB during a vote through other means, should they so choose.
2. Disallow LobbyFi and keep the current system, knowing full-well that shady backroom deals may still occur that achieve the same thing. Though to be fair, that kind of dealing surely still happens from time to time, even with LobbyFi.

(I recognize it’s not as easy as just “allowing users to accumulate more ARB during a vote”, preventing situations like passing around the same large chunk of ARB between several voters to temporarily inflate their voting power, etc, has to be accounted for on a technical level, but in my mind there’s no point in digging into implementation details before there’s alignment on whether it would even be the right philosophical approach)

[JoJo]

I am positively surprised that @Arbitrum has been starting the conversation on this topic.

I’ll post and discuss some of the points that I have been personally discussing (for months now) in the delegates chat.

Governance token and bribing markets

First, we should not ban or censor lobbify, or systems like lobbify. Crypto has been so far build on the premise of composability, and one of the perk we are seeing compared to traditional finance vehicle is that you can rent and give out your voting power in exchange for a payout. This has happened so far in convex, curve, balancer, and in general in any token-emission environment in which directing emissions is subject to a choice of the owner which can subsidize this choice and power to a third party.
What we are seeing in ARB is one of the first instances in which this is related not to emission of token, but shear governance voting power.

This small paragraph is about describing the nature of tokens having a governance power. Doesn’t matter the power they hold: emission, voting, both, or others, as long as token has a power in the ecosystem, that power will be rent to third parties that can optimize votes, create a yield for the owner etc. We should not fight the nature of crypto tokens with voting power.

Solution to improve the current situation comes down to 3 points:

1. increase the ARB voting power distributed to aligned delegates, and increase in general the appeal of delegating ARB to active delegates
2. specific tech changes to Lobbify
3. potential changes to our Code of Conduct

Before that is worth understanding what proposals are more impacted by any bribing market for the history of arbitrum. Useless to say, offchain elections: tally votes are usually a confirmation among a certain set of candidate with a binary yes/no so the effect of 20M votes is quite lowered. At the same time, weighted elections with shutters are were lobbify can impact the most through a single choice. This is were I think we should focus our efforts right now, I don’t currently see tally proposals being in danger. Not only we have the security council, but also, if a proposal on tally would pass by measily 20M votes, means that it was a contentious one from the start and Lobbify influence is less relevant than this contentious nature.

1. ARB voting power

We all know about the issues about stagnating voting power, which doesn’t circulate and get redistributed regardless efforts of AF and others. Few solutions here, already mentioned by others:

• having ARB native staking: however is financed, anything that is non 0 will create a realistic alternative to lobbify
• having an LST for ARB staking that has more rewards but also funnels voting power into “Arbitrum aligned parties”. This was already discussed and studied at a superficial level in November, and I personally worked on the governance part alongside @jameskbh. At the time we advocated for automatic strategies to redistribute governance voting power automatically in the DAO to certain delegates. It was sad at the time that the general reception was: this is too complex, we risk breaking thing, because the current result was foreseeable
• having ARB delegated directly from treasury to certain delegates in the DAO. This would be a temporary solution, doesn’t solve lobbify being a way to generate yield through delegated ARB.

Our goal should be for lobbigy, at 20-30-40M votes, to be a not so decisive part of active voters because we increase what delegates have.

2. Specific tech changes for lobbify

As it is now, even if I was participating in the OAT election, and even if I bought the votes for myself, I could have not voted, in an equally split way, for myself plus other 2 candidates. It means i could have not used lobbify votes in a “compliant” way even if I wanted to. Working on this would allow for a better usage of the platform. And yes, I know the vote can just come from an anonymous user, but we should in the first place find a way to use it for ourself if we want to. Paging @lobbyfi on this point, and we had a private discussion on how this is non trivial on their side.

3. Potential changes to our Code of Conduct

This is potentially tricky.
For this, or any voting platform, who would be the users in case of an election that would be incentivised to buy the votes?

1. people who are candidates in the election
2. users who are not candidate in the election, but that want to have an higher voting power to ensure a certain outcome.

For 1., is easy: if I buy the vote for myself, i should vote it in the same way i do it with my own delegated token. I can disclose I bought it, and that I voted myself, X and Y in an equal weight manner as per CoC. To have this, we need a change in the lobbify platform as stated before
For 2., is honestly a mess. The outcome that that specific party might want could be just an opinionated one (“i like X more than Y”) or the most disruptive one (“i think X is the less capable person in the pack, so I am going to vote for X to ensure the initiative has less chances to succed”). For people who say this won’t happen, it can.
Reading in the intention of a voter is, honestly, quite complex. A different opinion can easily be seen as unaligned. It might come down to the following:

Case A: for elections of several members at once: if whoever bought votes voted in a way that is compliant to the CoC (in the OAT case, 3 members, equally or around equally distributed), all is good
Case B: for elections of several members at once: if whoever bought votes voted in a way that is compliant only for non candidate, but we can’t source who the buyer was, we might have to disqualify the user, unless he/she specify in the forum his willing to express that vote in a certain way and declares to be non affiliated to the party. This is easier said than done and might require onchain sleuthing to validate the vote
Case C: for elections of a single member: CoC states you can vote for yourself.

Summing it up, case A and C are trivial, while case B is quite complex.
Who does the onchain sleuthing?
Who validates the results from sleuthing and takes a decision?
Who ensures that the entity that declared the vote stated the truth and is not maybe affiliated to one of the candidates instead? (and it might be very well the case)

These are open questions to which I don’t have a specific answer right now.

---

TLDR: most of the effort should, honestly, be put into increasing voting power. We should also try to work with lobbify to find the proper solution, eventually going to a case by case scenario in certain offchain votes like complex elections such as this one to ensure that there is at least the ability to be compliant in the usage of the tool.
Banning it, I am not sure would benefit anybody, would mean really going against of the main functionalities we have seen since the DeFi inception for governance tokens.

At the same time, I share the concerns of the Foundation. In the hypothetical scenario in which we had an election with one or more parties being elected with Lobbify, with a winning margin below the total amount Lobbify has as VP, we would always have had the question “did we get the best person appointed by the DAO”. Or even, “was the result piloted by someone not aligned with our DAO”.
We definitely need an honest discussion about this, and I am glad is coming from the Foundation.

@Arbitrum

Finally, it is never a boring day in the Arbitrum DAO and this is another big moment for us.

Neve a dull day in Arbitrum indeed

[jameskbh]

Thanks for putting this item into discussion.

I want to share a few comments regarding the following point:

@Arbitrum

Above all else, in DAOs, the goal should be to implement a governance process that can lead to positive outcomes for the community.

IMO, a good governance process is robust enough to work DESPITE attempts to break it, ensuring fairness in decision-making. The framing I’m making here is:

governance process = set of rules + engagement

If you have the rules but not enough engagement, it will not work. If no rules, and a lot of engagement, it will not work either. Hopefully, a good set of rules with healthy engagement will keep things on the right track.

It is not feasible to try to create a set of rules to prevent every possible offense to the process. This is the discussion we, as a DAO, have been having for some time in this particular case. However, I see the question presented here with a different wording:

How do we prevent bad engagement from happening?

The short answer is: we can’t. Under our current system (delegated), the way to fight this is to increase representation/active voting of good engagement. And that can only happen with more voting power being active in the decision-making process.

@ajwarner90

For those who don’t recall, when Arbitrum transitioned to a DAO and a decentralized protocol, the power that was transferred to the DAO was literally unprecedented. What made Arbitrum particularly notable was the onchain control of the protocol by tokenholders,

Only having more token holders engaged (directly or by delegation) will solve this issue. Vote buying is not the real issue; Voter (tokenholder) apathy is.

I won’t repeat the points @JoJo laid down in his comment, but I share the sentiment: It is mandatory to increase the overall engagement to dilute the weight of bad engagement (when/if it happens). The DAO started to explore some tools to deal with it a few months ago. I believe it is time to resume this conversation.

[dennison]

Hey folks, Dennison from Tally here.

The situation with LobbyFi and Vote buying is very much what we had in mind when we originally proposed Arb staking.

At it’s core the issue here is that someone with a large amount of ARB is dealing with the opportunity cost of holding ARB as an asset versus the putting their capital to work in some other way (Selling their ARB and buying gold).

I assume the owner of this 20m ARB believes in Arbitrum generally (they haven’t sold, have a huge position, and are massively positioned for upside), but is thinks enough about the capital inefficiency that they have decided to seek yield by selling their voting power.

LobbyFi has provided a way to alleviate the opportunity cost of holding a large amount of arb by selling the voting power. This is entirely allowed by the current system and indeed is probably impossible to prevent (they could always sell their voting power in other ways and we would never know). By using LobbyFi in many ways they are effectively making a statement as a shareholder in such a way that potentially endangers the system. (I would assume they probably trust the security council to intervene if there were any true danger as a result of selling their vote).

Truthfully I think the way you address vote selling is address the root cause of opportunity cost of capital, and hence, the idea of Arb staking. Obviously thats a much larger discussion but its worth thinking about the inspiration there: how can we get to an ideal state using incentives? What could be done so that vote buying at scale like this just wouldn’t exist? Is that possible?

[Vertex_Protocol]

@EzR3aL

What if we as a DAO could vote upfront in a LoobyFi thread for a price range? Because if the DAO agreed to the price, then its a fair mechanism.

@JoJo

2. Specific tech changes for lobbify

@JoJo

We should also try to work with lobbify to find the proper solution, eventually going to a case by case scenario in certain offchain votes like complex elections such as this one to ensure that there is at least the ability to be compliant in the usage of the tool.

LobbyFi might be willing to negotiate things like this, but the next project that gets set up would have no obligation to and it’s a competitive disadvantage if your pricing and development goals are set by the DAO while others have the freedom to do what they want. If LobbyFi wants to work with the DAO that would be great, but we think it’d also be fair for them to refuse to alter their development priorities, protocol architecture, or business model to fit with our preferences.

@jameskbh

IMO, a good governance process is robust enough to work DESPITE attempts to break it, ensuring fairness in decision-making.

Also will not repeat all the points from above but we strongly believe that this is the view the DAO should be taking. Sooner or later there will be an actively malicious entity that acts in the ecosystem, and they don’t have to do something as dramatic as swinging an election by a small margin. If you are [insert your least favorite chain here] that wants to slow Arbitrum down as much as possible, it would be as simple as voting against any amount of money being spent on programs like STEP, D.A.O., DIP, or voting for on things that you believe will drain the treasury for little impact.

[JoJo]

@Vertex_Protocol

LobbyFi might be willing to negotiate things like this, but the next project that gets set up would have no obligation to and it’s a competitive disadvantage if your pricing and development goals are set by the DAO while others have the freedom to do what they want. If LobbyFi wants to work with the DAO that would be great, but we think it’d also be fair for them to refuse to alter their development priorities, protocol architecture, or business model to fit with our preferences.

For sure lobbyfi doesn’t need to agree with the DAO or even answer our questions. But Denys showed to be quite open to discussion on how lobbyfi can coexists in our ecosystem in a productive way, both in the delegates’ chat and in private, so I think is worth to try and find a common ground here.

For new competitors that could literally fork lobbyfi and not comply to any rule: my educated guess is that whatever stance we get through lobbyfi, would not be an ad-hoc one but a generalised one. If a new platform comes and doesn’t comply, after reaching out, we might have to unfortunately not count the vote on offchain votes. And yes offchain votes are just one part of the equation in a DAO with a treasury all on chain, which pushes us back to the initial statement of improving resilience and voting power of the DAO as a whole.

[cypherbadger]

In my journey observing how governance systems emerge, evolve, and sometimes disappear, I’ve seen it all.

Back in the days of DAOs operating as proof of work systems with shared treasuries, rewards were distributed among contributors. Deals were made with masternodes, ensuring they continued operations with the expectation of receiving a significant share from every proposal that passed.

The difference now is that all of this can be made programmatic, trackable, and auditable — until zero knowledge technology fully shields us.

So what will we do when voting and vote-buying become completely private? It is technologically possible, and future governance models will likely embrace it. What happens when the voter — or the one executing the power of a vote — is an AI with enough autonomy and capability?

Are we going to sit down and debate how to ban it?

To me, the answer is to adapt and create mechanisms resilient to these vectors. We don’t want to suppress vote deployment through various creative strategies. What we must safeguard is that, regardless of how votes are obtained or exercised in the system, the ecosystem and the network must ultimately benefit.

This is the principle of antifragility — something we inherited from Bitcoin and something I believe is worth protecting.

Privacy matters, yes. Permissionless voting matters, yes. And designing incentives for all these novel dynamics should also matter to us. This is the magic of composability at the end of the day.

Let’s be an antifragile community — not just say we are. The tools we use should reflect that antifragility too.

[EzR3aL]

@Vertex_Protocol

LobbyFi might be willing to negotiate things like this, but the next project that gets set up would have no obligation to and it’s a competitive disadvantage if your pricing and development goals are set by the DAO while others have the freedom to do what they want. If LobbyFi wants to work with the DAO that would be great, but we think it’d also be fair for them to refuse to alter their development priorities, protocol architecture, or business model to fit with our preferences.

My best guess is that forks of LobbyFi or similiar competitor would try to adapt the approach we and LobbyFi will have in the hopefully near future.
Lets say they do not want to work with us and we block them again and again, this would harm their userbase and the user experience and they simply leave for LobbyFi then.

But of course they can simply say not, LobbyFi and any other competitor. But so far I have only witnessed that LobbyFi is willing to find the best way to work with Arbitrum and we should do as well, cause no matter we don’t like buying votes, the mechanics are there and they aren’t going anywhere.

[TodayInDeFi]

Not a fan of vote buying but also dont see how it can be totally prevented long term, especially as privacy technologies develop further.

One thing that could help mitigate it greatly would be accererating the implementation of ARB staking.

Without staking or revenue share to ARB, the only valuable use case it has is to stake with Lobby Fi

Providing an alternative yield on ARB would compete with LobbyFi on all but the most competitive proposals, reducing this issue greatly.

[pedrob]

Thank you very much for starting such an interesting discussion.

I mostly agree with the sentiment expressed by the delegates who have shared their views. In general, I don’t think it’s a very feasible strategy to try to fight against market dynamics in a system where voting power is directly tied to economic value—especially when that voting power can be acquired in various ways.

That’s why I agree that censoring @lobbyfi is neither the solution nor the right response to its existence. I believe @olimpio expressed this perfectly:

@olimpio

In open, permissionless systems, whoever can create an incentive or have an edge will do it, and it should not be censored. If attempted to be censored, new ways will be created, and it will become a cat a mouse game. It’s not like this is entirely new - governance tokens can always be borrowed before a proposal, delegated, and used to vote. And lobbying already exists in the private setting.

Instead of punishing those who have found a way to give idle ARB economic value and utility—beyond simply delegating it, which currently offers no direct benefit (aside from the potential long-term upside of supporting a high-performing delegate)—I believe the DAO should focus its resources on aligning incentives toward behaviors that are considered most beneficial in the long run.

And this goes hand in hand, as @JoJo mentioned, with the development of ARB staking. What we can start doing is revisiting the discussions around what kind of yield will be offered and where it will come from.

Now, the other discussion surrounding vote buying relates to the level of coordination required to maintain the DAO’s sovereignty over Arbitrum—both in terms of the treasury and the technology. And how we actually agree on “what the DAO wants.”

Would there have been a feeling that “this isn’t what the DAO wants” if @cupojoseph ended up in the top 3? I’m not sure. One could easily argue that the answer should be “no”—because the whale who put their voting power up for sale is also part of the DAO by virtue of being a token holder, and in doing so, it expressed its opinion.

Does such a thing as “what the DAO wants” really exist? Or is it even possible to reach a consensus on that? The SOS process seems to be an attempt to explore that question.

Should we just let “market dynamics” determine who gets to sit in a position that can have a very direct impact on the future of the DAO? What’s the potential risk for Arbitrum in accepting that dynamic?

Unfortunately, I don’t have all the answers to these questions.

But perhaps, as a starting point, we can acknowledge the existence of the dynamic and begin designing mechanisms to prevent outcomes that may not truly represent the “majority consensus” — whatever that may mean.

In the end, it all comes back to incentives alignment.

Looking at the specific case of the recent election—was ranked and shuttered choice the best way to go about it? Or could there be another mechanism that not only reflects the actual distribution of voting power at that point in time, but also prevents a sudden and isolated move from having an outsized impact that could disrupt long-term planning for the DAO?

If we take that potential impact into account, we can design mechanisms that either minimize it—or ensure that, when it does happen, the move is significant enough that there’s no doubt it reflects the will of the DAO.

[web3citizenxyz]

Thanks for facilitating this discussion. Our two cents on Lobbyfi and voting-as-a-service. I think there’s two main questions:

Can we get bad outcomes having vote buying services?
I think it’s obvious that it can have bad outcomes in vote buying . So its more productive to focus on the second question.

a simple example

Hiding this example in the toggle to keep it shorter.

Say some individual person nominated themselves to become a OpCo OAT member, disclosing no financial, personal or professional COI and complying with the KYC agreement. Their motivation simply to generate adverse effects in governance by occupation of a key role and not fulfilling its mandate. Then, asking someone to buy the 19.2M votes from Lobbyfi and casting a full vote for them. It is obvious then we have a bad outcome facilitated by vote buying.

An example like the above one can be facilitated without the existence of a platform like lobbyfi by simply applying pressure/agreeing/bribing a large delegator to vote for yourself .

Can we control outcomes? Do something to mitigate/prevent it?
As many have said: we can’t control the existence of vote buying markets and banning won’t control it either. So we would not support this. Also we can’t control the external personal motivators driving the buy/vote mechanism.

A more productive discussion can be how can we can better engage with Lobbyfi and how to better align incentives.

These are more general comments as we don’t have a formed opinion on these:

• Engage with lobbyfi “embracing it as a part of governance” as paulo stated above and suggesting a better way for pricing votes to prevent governance capture when it applies. We can’t control pricing but Lobbyfi have been active in the chat and in the forum, maybe we can suggest other improvements beyond pricing.

• It would be interesting to assess the outcomes of the ARDC report if it has some recommendations. But also we should be looking into what other economic incentives can we present to those simply looking to generate yield for ARB other than staking.

• Ideally all delegates would engage with governance. How can we engage with large holders to make them more active or delegate their vp? Can we explore delegation to active delegates increasing representation of delegates engaging with governance?