Meiji Governance Rework

meiji

Authored by : @ControlCplusControlV

This proposal is the sequel to the initial Meiji Restoration post, and includes many updates. The Governance rework of Meiji is also being separated out to make it both more focused, and allow for tokenomics/governance changes to delay each other.

The Sushiswap Meiji DAO will supersede all responsibilites currently held by the Sushi DAO. The Meiji DAO will bring governance on chain, and kickstart the Meiji Restoration of Sushi, which will become a new phase of Sushi and a new grand vision to execute on.

The New Meiji DAO

The Meiji DAO will require shares instead of tokens, and feature a new quadratic voting mechanism. The Meiji DAO will also be fully autonomous on chain, and allow the Sushi DAO to explore a new era of on-chain governance and operations.

Membership and Shares

To prevent Shares from just becoming tokens again, they are granted by locking up SUSHI in an entrance contract (of which you can exit at any time, forfeiting shares), and follow the formula of Shares granted = sqrt(Total SUSHI you've locked up) / log(sqrt(Total Members)). This formula was chosen because it results in a distribution similar to that of Quadratic voting, where whales see diminishing returns in shares granted instead of emphasizing the power of the average contributor, but the denominator slows the effectiveness of Sybilers. Additionally, the Meiji DAO is protected from Sybilers since a Member can be kicked out of the DAO for malicious behavior (not disagreements, but Sybiling and other destructive behavior) with a vote where at least 80% are in favor.

This Mechanism is shown below, at different member capacities. (The number in the log(sqrt (_)) in the denominator is how many members are in the DAO, with the x-axis being SUSHI locked up, and y being shares granted)

The tables below show shares granted at various thresholds of members and Sushi locked up into the DAO. The result is a mechanism that has some in-built Sybil resistance, which compliments nicely with a minimum share threshold and member slashing to protect the DAO from Sybil or other governance attacks.

Total Members 100 Sushi 5000 Sushi 10000 Sushi
100 10 shares 70 shares 100 shares
1,000 6 shares 47 shares 66 shares
60,000 4 shares 29 shares 41 shares
1,000,000 3 shares 23 shares 33 shares

Shares are also not granted immediately, and instead, Members must wait for double the voting period before they can vote on proposals. This is designed to prevent new voters from joining the DAO in an attempt to attack or vote through a specific proposal without prior service or participation.

High Kitchen

As a final backstop against sybiling and other attacks, an optional “High Kitchen” can be instated made up of trusted community members and contributors. While not granted special voting powers, this “High Kitchen” can veto proposals through the DAO with the intent of this power being used to defend the protocol against governance attacks, or similar vectors which arise from the new DAO being on chain.

It is important to note this High Kitchen will be chosen by the community on the finalization. The High Kitchen cannot rush through any proposal, nor do they have the power to act on behalf of the DAO. They can merely block proposals from being executed if they are deemed harmful, and a last line of attack.

The High Kitchen will only have the ability to veto a proposal so long as they achieve a majority. They can also vote to dissolve the High Kitchen with a majority vote.

The suggested high kitchen members are currently suggested to be the Treasury Multisigners, as those people currently serve a similar purpose of protecting Sushi’s assets.

DAO Operations and Voting Process

The main parameters for voting are as follows -

Vote Duration : 7 Days

Entrance Duration : 14 Days

Quorum : Participation of more than 5% of Voters AND 10% of shares

Minimum Entrance Bond : 1 SUSHI

Signal threshold: Top 3 proposals each week, above 1% of total vote.

Minimum to post a proposal: 500 Sushi in shares or a combined group equaling 500 Sushi in shares

Signal Phase

Proposals are a multi-call transaction which can be suggested by anyone above the post minimum requirement. Proposals then sit in the signal phase for up to 2 weeks, or until they pass the Signal threshold. The top 3 proposals above threshold each week (ranked by signal strength) go up for execution voting.

Execution Voting

Once a proposal enters execution voting, then voting begins for the vote duration of 7 days. During this phase, for a proposal to pass it must recieve a majority of votes and recieve a total amount of votes above the Quorum threshold. From there the proposal will enter execution.

Execution

Once a proposal passes it can then be executed by function call on the Meiji DAO, which will execute the proposal. For a proposal to be executed, it must be executed in the order they are passed. Proposals in the execution queue however can be voided with a seperate proposal.

Easy Track

While voting can be contentious, some things should not be. Proposals like backpay, renewal of agreements, can often go to a no-contest vote where the signal change is unchanged, however quorum is reduced to 50% of original quorum so long as no more than 15% of the votes in weight or voters are against the proposal, in which case the proposal fails.

The Sushi Org’s Future under Meiji

The Meiji DAO will take up all responsibilites of the previous Sushi DAO. This means paying team salaries, controlling treasury, and funding expansion. Infrastructure will also integrate Furo to allow payment of salaries directly as an action.

Implementation

The Implementation of Meiji will be led by ControlCplusControlV, with assistance from the Sushi team. Once audited the Meiji DAO will then be deployed, and a final vote will transition over governance powers.

  • Yes
  • No

0 voters

7 Likes

Does this mean that anyone who is able to sybil a >20% share can only be kicked out by the High kitchen? Presumably it is very difficult to detect sybil activity. For example:
image

Also, how to think about second order sybil’ing, such as vote bribing? If I spin up a bribing protocol tomorrow, I can bribe smaller holders to vote in my favor, since the “first” xSushi is worth the most in terms of voting power. Would participating in a bribing protocol be grounds for removal? I feel like the “removal” point will be contentious, in terms of what should define removal.

3 Likes

Yeah, perhaps the ratios and parameters should be adjusted from lower than 80%, and redelagate this power to the high kitchen, or perhaps just a simple majority.

I would consider vote bribing sybil behavior since it’s anti-democratic governance, but I also agree this should be more strongly defined, did you have anything in mind?

2 Likes

Yeah I guess the question here is a first principles question:

  • Do you just accept the existence of Sybil behavior as a given attribute of DAOs and design something that is indifferent to Sybil-ing (Token voting is of course one such example)
  • Do you try to actively work against Sybil behavior with mechanisms like the above, while potentially introducing new points of contention (bribing, “removal” criteria, etc)?

I personally like the spirit of sybil-resistant voting that doesn’t give undue influence to large holders but just trying to point out the tradeoff that’s being made in design

obviously by definition any function which has a decreasing marginal vote power is subject to sybil pressure

1 Like

You can catch a lot of bees if you put a bag around their nest imo

2 Likes

I think it’s worth making the lives of Sybiler’s miserable, without being ignorant to their existence. So I think we are on the same page about this. At it’s core, the DAO should slow Sybiling on the technical layer to give time for the social layer to respond

2 Likes

I had someone ask me questions about Meiji, posting my response here


  • So Entrance Duration is the time from depositing Sushi to getting them as Shares, to prevent flooding before a vote. So they can only post a signal if they already have enough shares to do so, or wait until after there shares are granted

  • Proposals aren’t affected by the entrance duration, once they reach signal their 7 day timer starts for voting

  • Voting for execution is non-competitive, so once a proposal is in the execution phase it must simply meet the requirements to be executed.

  • Shares just mean votes/(voting weight) in the DAO. The requirement is just an initial lockup of 500 Sushi, or a cumulative total locked Sushi in the DAO of atleast 500 Sushi.

Execution Voting:

  • For a proposal to pass it must clear a majority vote and quorum, but it doensn’t need to need to best the others, competition only occurs at the Signal phase. That would be very interesting to add though and autorenewal atleast once if it doesn’t pass quorum but gets majority is worth considering
  1. One thing I think would be valuable for the community/meiji members is having a required structure for signal phase proposals to help consistency and clarity. Something along the lines of ARC’s can ensure proper proposal hygiene ARCs - Governance](ARCs - Governance)
  • I agree on Signal, although its not always necessary, I wonder if this a good way to optionally include that? (Maybe tx’s that use treasury funds?)

Voter fatigue I mainly thought by limiting the number of concurrent proposals to be voted on. A penalty to make it harder to get signal for a proposal if you frequently post unsuccessfuly proposals would be very useful.

4 Likes

Congrats @ControlCplusControlV on getting this proposal reworked & for giving (selflessly) so much focus on revitalizing & energizing the protocol. The community is lucky to have you. It was great to read through these ideas and consider some of the longer-term impacts, risks, and benefits over the last week or so

Below I’m including some comments, suggestions, ideas, and questions - hopefully, this helps get to a constructive accommodation of feedback to make this proposed implementation EVEN stronger & setup SUSHI to have a vibrant, engaged, and active community of regular participators

Issue with choice of passive income vs governance rights
As things currently stand, SUSHI token holders wishing to participate in Governance also gain passive income benefits by gaining fees either via xSUSHI staking or through providing liquidity to the SUSHI/ETH pool.

The proposal above effectively fully separates Governance from economic benefits & forces SUSHI holders to make a choice between staking for fees vs creating shares for voting rights - whilst fully acknowledging that SUSHI holders can always unwrap their SUSHI and burn their Shares this does mean that they will be limited in fee-earning potential all the time an individual wants to participate in Governance. In simple terms, this presents a dilemma as it will make participation for some less attractive than under the current mechanism - for existing SUSHIPOWAH entitled holders this will create a sense that they have lost one-half of the features/benefits they’ve been used to receiving

Potential Solve
This dilemma can be overcome if two accommodations during the initial implementation can be considered;

  1. Initial shares are granted to all current SUSHIPOWAH entitled holders
  2. Shares implement a ‘decay’ feature that erodes some value of the Shares power for non-participation over a relevant time period

I would also suggest it would be good practice to utilize the decayed shares to reward active participants in the Protocols governance by the High Kitchen having the power to grant new shares to valued members of the community that have contributed in significant or meaningful ways (examples might be ControlC+V successfully implementing this proposal) - more on this power below

Once past the initial implementation, it would be fair, equitable, and sensible to then switch to the new mechanism with a choice between either.

Goal of Increased Participation
The role of governance is more complicated than simple voting and is inclusive of the ideation, forum and discord debate, research & votes that are generated by a vibrant and active community.

Luckily the forum already has some datasets that allow for understanding the quality of contributions that serve as a good basis for further enhancements. Where/if there is no real reward or incentive available to contributors that drive these efforts then participation will continue to be a risk that the protocol has to navigate - this risk will get greater if/where governance gets centralized by participants that favor owning governance rights over fees entitlement (especially where the suggestion above is NOT implemented or considered in full)

Potential Solve
This can be solved through the implementation of;

  • High Kitchen having the power to grant governance shares to meaningful contributions to community members
    • (in concept this can work similar to the feature implemented by MolochDAO whereby Shares that were not created through wrapping SUSHI have no future claim on any underlying assets - see Shares vs Loot from MolochDAO docs)
  • Refining the existing forum structures so that Share holders also have a points system that can be used to upvote or downvote meaningful contributions, comments or proposals - thus determining new share grants that must be ratified or issued through the High Kitchen

These solutions could be implemented in phases, with initial phase being under full control of High Kitchen then the second point implemented at a later stage as a further improvement of the mechanisms

Securing the Role & Goals of the High Kitchen
Currently, the role of the High Kitchen is described at a low level of detail. Fleshing this out so that there is greater clarity on the;

  • Role of High Kitchen
  • Nomination, Addition & Removal of members (vs just utilizing treasury multisig)
  • Scenarios in which High Kitchen can be dissolved + then what?

Potential Solve
This can be solved by;

  • Put options on the role, rules, and powers of the High Kitchen to vote
  • Put the selection & nomination of High Kitchen members to a vote

I would also recommend that one of the first duties of the High Kitchen should be to formalize a new DAO Constitution for Sushi leveraging information like the great work metagov has done

Final Note - on the concept of decaying share rights
Desire here is to clarify that any decayed share rights as proposed above would not affect holders’ claim on underlying SUSHI & that conditions for decay should be fully considered and debated. A good starting point may be as simple as;

  • a % loss in total share power if a voter hasn’t participated in any votes within a defined period (i.e. did not vote on any proposal within a 3-month period - only applicable where there were votes within that period)
  • a % loss in total share power if the holder didn’t participate in comments or signaling votes for Proposals during a longer set period (i.e. over a 6-month period)

Of course, this would require some dev efforts to grant the High Kitchen multisig the power to process decay commands to the shares but should be possible. The effort to implement would likely be justified if it leads to stronger governance participation than in current implementation & can always be revisited if the benefits do not materialize after a suitable period (i.e.18 months after implementation)

2 Likes

Issue with choice of passive income vs governance rights
As things currently stand, SUSHI token holders wishing to participate in Governance also gain passive income benefits by gaining fees either via xSUSHI staking or through providing liquidity to the SUSHI/ETH pool.

Yeah I agree this isn’t ideal. Perhaps from just pure inflationary Sushi SUSHI/ETH LP could also be allowed, or another form of value-accruing Sushi.

I do also like the High Kitchen to be able to bestow shares, maybe only in chunks of “awardable” shares as voted on by holders to prevent takeover.

The role of governance is more complicated than simple voting and is inclusive of the ideation, forum and discord debate, research & votes that are generated by a vibrant and active community.

I like the Loot vs Shares distinctions, and as mentioned above allowing the community to allow discretionary votes to be allocated by the High Kitchen. Integrating with the existing forum is something I do also want to explore more but it was too much to fit into this single proposal. But I agree a staged approach with this would be very cool, and could help incentivize participation.

I agree with your statements on fleshing out the High Kitchen vs reusing the Treasury Multisig. It looks like a seperate system will be needed to elect those, so that will be come another process to select them after Meiji Signal.

I will have to think about decaying share rights because that would require a lot of changes to the backend side, so I’m not sure what that would look like but some activity multiplier would be useful to weight more active community members higher

1 Like

Yeah, the decaying method would be most applicable where current xSUSHI and SUSHI/ETH LPers are automatically receiving some Shares (perhaps via an airdrop with no “loot”/underlying) & that can be complemented by new SUSHI holders wrapping for Shares

This would be my recommendation & best outcome because it solves the dilemma I described but also then effectively removes non-participants (or they could delegate their POWAH to someone else within the community) if they want Share power to be upheld over time

If it goes the other way, where proactive activity causes a multiplier, then that would likely make it less of an incentive on the “awarded” shares idea - that’s because those shares would start from a non-multiplier, so would have less POWAH. Whereas if the awarded shares are active against other shares that are decaying then they have more value to the holder from a governance perspective

I guess the other way to think about it would be to automatically apply a multiplier to awarded shares, so that they have strong value from time of award - but that will likely be perceived as less-fair and has some risks of favoritism & controversy should the awarding mechanism not be really really clean & transparent

Also wanted to say - although I do believe this would be a big improvement on the governance structure as a whole, I’d really want to see some further revision of this rework to incorporate some of these factors before I’d be happy to flip over to a “Yes” vote

Really, the work here is fantastic. I think though given the significance that it will have on the future culture and workings of the DAO, if implemented, then it would massively benefit the proposal for it to be tightened up further to factor for these considerations - let me know if any of the feedback needs further clarifying

2 Likes
  • Do you just accept the existence of Sybil behavior as a given attribute of DAOs and design something that is indifferent to Sybil-ing (Token voting is of course one such example)

Yes.

My question is does Decay encourage activity for the sake of activity and maybe lead to purposeless voting? I think a decaying multiplier would be good to add to awarded shares as well. With the community voting on the High Kitchens Shares allocation power I don’t think it could become too corrupt as the community would still have that check over what they can allot

It’s a good point, maybe a crude workaround for that is to include “Abstain” by default on polling & votes?

100% agree - rule should be consistent for all.

Yeah, makes perfect sense as a check & balance of powers.

1 Like

On this one, I was thinking about the point of the dilemma I mentioned faced by users (i.e. Earning fees vs Governance) and on-balance I think you should just airdrop the initial governance shares to anyone that currently holds either of the SUSHIPOWAH assets. Then for anyone who wants to get into governance after that initial airdrop they have two options;

  1. Buy SUSHI & wrap for the Shares (with Sushi getting locked as “loot”)
  2. Earn through High-Kitchen allowance awards, etc.

If there is enough notice ahead of this happening then it’s likely additional SUSHI demand would happen for the SUSHIPOWAH ahead of airdrop - so a net positive impact on price in the short term.

I support moving toward a voting model akin to quadratic voting—the system you describe seems like a move in the right direction.

@Funk raises some valid concerns, especially with respect to separating passive income from governance rights. Requiring voters to give up financial benefit in order to vote creates an opportunity cost to voting that will make it less likely for everyday users to dedicate their tokens to governance. This could lead to increased influence for whales which seems contrary to the point of the quadratic model.

I’m sympathetic to the idea that those who actively participate in governance by voting should have more relative influence. However, I’m skeptical of giving governance rights to users who are engaged in the forums and driving proposal creation. This results in a feedback loop where the most active accumulate more power even though they do not represent the average protocol user. It’s reasonable to ask people to participate in voting in order to maintain their governance power, but I’m not sure governance participants should have their governance power diluted because they aren’t participating in the forums or submitting proposals.

This is my first post in the Sushi forum, but I recently joined the Flipside Governance team, so you’ll be seeing me around. I’m looking forward to doing Sushi governance with you all!

2 Likes

Well put, exactly the point I was trying to make. Most likely outcome in this scenario is that smaller SUSHIPOWAH holders would choose to opt-out of the governance shares creation.

For that reason, I favor managing the initial issue of governance shares via a snapshot of all SUSHIPOWAH holders. If announced in advance (i.e. today + 30 days) then anyone not currently a SUSHIPOWAH holder would still have time to opt-in and receive the initial issuance of shares. Once past that initial period the only way to get new shares would be in the 2 methods described in my post above

It’s definitely a fair consideration, but also one that is somewhat mitigated by the role that the High-Kitchen could play in determining the relative value of governance contributions, especially where paired with a new shares issuance schedule (this could be based on a decay factor or a new share release schedule - I favor the decay)

This is a good point to clarify on;

  • non-participation in voting = candidate for decay (ideally with some limit to non-participation over a set/fixed period - for, example, if shares holder doesn’t vote on 5 proposals in a row, or less than 3/4 of votes in a 12-month period, etc.)
  • shareholders wouldn’t suffer any decay for non-forum or non-proposal creation - instead this would be focused on incentivizing participation through new share awards (without claim on any underlaying SUSHI)

The two, together, act like opposing forces of incentive / disincentive

Welcome :wave:

1 Like

Thanks for the welcome and clarifications. :cowboy_hat_face:

Regarding the incentive/disincentive issue, I understand the distinction you’re making but it doesn’t feel relevant to majority-rule governance. What matters is a participant’s relative governance power, so giving one person more governance power (shares) is the same as taking a little bit of governance power away from every other shareholder.

Issuing shares to forum contributors lessens the relative governance power of other shareholders even if shares are issued to contributors rather than using a decay mechanism. Allowing the High Kitchen to value and reward proposals with governance power will likely lead to a concentration of that power.

NP, love the healthy debate and that’s exactly the kinda feedback that should help strengthen the proposal overall - so very much appreciated :kissing_heart:

I think that’s a bit dependent on tightening up the design, no? Decayed total amount vs newly issued shares doesn’t necessarily need to be equal to 1:1. The newly issued shares (less any “loot”) can be based on a constant emissions rate or emissions curve that can only be changed through further governance & we should also consider that new shares could be created at any future point where SUSHI is wrapped for shares. The only hard limitation on that is max supply of SUSHI since theoretically, as improbable as it’s likely to be, all SUSHI could be wrapped as shares.

It is also possible that awarded shares (through this means) could decay at a faster rate than non-wrapped SUSHI shares - although it likely makes implementing the whole thing more complex than it needs to be.

I think this is also an area that’s highly dependent on its design. In the case that just ALL participation is rewarded then yes, it’s a fair point.

What I’d be keener to advocate for is focusing on the call-out that meaningful contributions are the ones rewarded - the example above whereby a proposal is successfully implemented could, for example, be designed so that proposer gets some amount of governance shares for;

  • Proposal being accepted
  • More when successfully implemented
  • Reflectively after a suitable period, where the change can be assessed as having been successful or not

Or, for example, where feedback on an initial proposal draft helped make it more successful, ultimately.

That would naturally reward and encourage community contributions and align those that care about the protocol’s overall successes & iteration with significant skin in the game.

Conversely, this contributor would be offset by others that have not added anything meaningful &/or perhaps not even participated in voting during a fair time frame. Those shareholders, if they had “loot” backing their shares, would still be able to remain their SUSHI & therefore still have some skin in the game since successful proposals can reasonably be expected to lift the overall value and worth of the protocols native token.

That being said, it’s probably a good time to clarify; I wouldn’t be in favor of decaying anyone’s shares based solely on non-participation in the forum or on forum debates. This should only be limited to non-participation in voting repeatedly.

It would however, be really cool that if, in the scenario, for whatever reason, a share holder didn’t vote then any decay (especially where “abstain” options are included on votes) could be off-set if they also made good/strong contributions to the debate on those proposals (although not really strictly necessary & probably overkill)

on abstain and delegate features

Delegation - delegation should be the baselayer for voting. easier to form(and see) power blocs and gain consensus on long term vision and platform and makes it so only a few people have to LARP that they’re across every possible facet of the org.
on that note, it’s waaayy geeky - but I’d love to think through “portfolios”/categories for proposals fleshed out and all proposals having to be assigned a portfolio - then us normie users can select their own “cabinet” to delegate their votes to on specifics.

Abstain - rather than being the default action ‘Abstain’ is a third, active choice. Signalling you want a solution/resolution to the proposal, but either don’t know or don’t care who wins. your vote power booster (or whatever you call it) doesn’t decay, and your action provides quorum for the proposal. you should be rewarded/incentivised for being engaged without it forcing arbitrary decision making
(the trolly problem but with a third option to swap with someone that actually want’s to pull the lever - you agree someone has to make the call and you can at least be like “well, they said they knew what they were doing”)

other possibly relevant crossovers from democracy stuff:

  • a “max vote decay” on proposals. the amount that you can contribute to a vote begins declining, starting at the max shares of a single delegate/wallet. forces whales to have to signal their intentions for proposals early and stay engaged to maximise their powah & shows smaller wallets/opposition voices whether or not their vote is even going to matter.

  • depending if we do the 3 proposals per week plan , we could fw an “attention market” for those votes. 3 proposals are up, if you don’t care and abstain from one or two your abstained votes go into pools and you’re rewarded with a share of that weeks “don’t care” pool. you can use the weeks don’t care points to get your share of the votes from the “abstain pool” of the proposal you do care about. but dis sum high level geek shit and i mean we’re not plotting the revolution here… or are we?

2 Likes