Current DAOs mainly rely on token based voting and delegation strategies. However, there’s been a lot of discussions around why token based governance isn’t ideal.
We would like to propose an experiment in governance at Rarible DAO which is similar to Representative Democracy of many western countries. The management of funds will remain in a multisig until the experiment is over, or, we decide to move forward with moving this fully on-chain.
Bullet Point Overview
People can submit themselves to become a Delegate, at any time. New Delegates may be added weekly or monthly (tbd).
Delegates are voted in and approved by existing Delegates (starting with a single member). Delegates receive a Delegate NFT which gives them certain rights.
$RARI token holders can then delegate their $RARI to approved Delegates.
The group of Delegates will have equal voting power (1 member, 1 vote).
In order for a Delegate to have 1 vote, they must: a) have a Delegate NFT, and b) have a minimum amount of $RARI staked on them
Delegates will have a list of requirements in order to remain a Delegate (such as attending a minimum number of meetings, or participating in forums and Discord).
All proposals would be approved by the Delegates via Snapshot.
We may decide to have a target breakdown of Delegate members, such as the diagram below, representing important demographics in our ecosystem. These percentages could change over time as the needs of the ecosystem evolve.
I think this is a very intriguing approach to representational governance X liquid democracy. Super cool idea.
A few questions to further explore this concept:
If a delegate only needs a minimum staked $RARI to be active and 1 delegate = 1 vote, whats the point of “over-staking” a delegate? Maybe we consider allowing delegates to transitively stake other delegates, allowing less-involved token holders to stake a lot of RARI on a single individual that they believe will make good decisions on how to distribute power. Almost a form of ad-hoc political parties.
Are the delegate NFTs transferable? It may be interesting, but it likely would cause problems. If the staked RARI are associated with an address, transferring the NFT won’t accomplish much unless the new holder manages to have RARI staked for their address. If the stake is associated with the NFT, the transfer of the NFT would be an effective unilateral transfer of power.
How many delegates are anticipated? Even several 10s of delegates seems like it may become daunting for the average token holder to sort through. The transient delegation idea above may help here to reduce token-holder burden and decision fatigue.
What are the details around new delegate onboarding? Is there a quorum % of YES votes required from existing delegates? Optimistic governance (onboard happens automatically unless its challenged) to allow more rapid expansion of delegates? Definitely some interesting things to think about here
I think delegation makes sense since voter apathy is real
“Delegates are voted in and approved by existing Delegates” → maybe also with a minimum % of tokens? Don’t want this to turn into an exclusive club
“b) have a minimum amount of $RARI staked on them” → what if I want Vitalik as a signer but he doesn’t have any RARI? should more stake = more voting power?
If you do with weightings, you could also apply multipliers to get the aggregate voting power of each group to arrive at ~33% team, 33% community RARI holders, 33% investors…or whatever mix you feel makes sense
Other DAOs have been thinking about this model as well and have some great content in their forums that I’d recommend to read through:
some comments from chat with Alex:
-how do you ensure alignment of delegates? Maybe they receive RARI token as part of this role
-he does like like the fact that delegates can remove from list and add to list at will. Maybe token holders can add and remove instead of delegates
I would personally (Eric) be OK with that second point, but think we should start with delegates being able to create the whitelist, and move towards giving this power to token holders. The main reason for that is that I do not trust the current token holders decision making ability, and think the group coming out of the centralized delegate group would be much stronger.
Okay, I put a bunch of thinking into this, and here’s where I’m at with it:
Design goals
First, it’s helpful to me to keep in mind the problems that motivate us to try an experiment like this in the first place. The current system, raw token voting, is 1) plutocratic, 2) vulnerable to vote-selling, and 3) vulnerable to low participation / voter apathy. So for new experiments, I think we want to:
Mitigate plutocracy: increase governance representation for important non-financial stakeholders like users, creators, and builders
Mitigate vote-selling: make it difficult for people who hold voting power to sell it to the highest bidder
Mitigate voter apathy: represent people accurately without requiring them to spend long hours engaging with governance
Maintain RARI importance: though these designs move away from pure coin voting, RARI token should still play a key role
Issues with the OP proposal
With those goals in mind, I think Eric’s idea might have a few issues:
Plutocracy & maintaining RARI importance
How much RARI does a delegate need staked in order to vote? If it’s a lot, that seems likely to be just as whale-dominated. If not…
Who gets to vote? If it’s only delegates, and delegates don’t need huge amounts of RARI staked on them, doesn’t that screw RARI holders? If both delegates and RARI holders can vote, what’s the breakdown of voting power between them? Who has more power over the creation of new delegates?
Voter apathy
If delegates are a constantly expanding group and they operate on 1-person-1-vote, won’t voter apathy among delegates be a problem?
I’ve got two variations to propose that might be a little better:
Option 1: “pRARI”
A new non-transferable token pRARI (“permanent RARI”?)
New contribute-to-earn program for pRARI, very similar to RARI liquidity mining (there can and should be overlap), except you get it for exclusively non-financial contributions (using products, creating NFTs, building for the DAO, etc.)
pRARI is exactly like RARI in terms of governance power, except you don’t have to lock it (it’s perma-locked)
You can vote with it
You can delegate it (not possible w/ RARI today, but something we can do and has been proposed in the past, I believe?)
emission rate of pRARI is designed so pRARI total supply will rival RARI total supply (perhaps exceed it, since whales are much less likely)
How does it do on the design goals?
Mitigate plutocracy: with pRARI being non-transferable, going to non-financial contributors, and eventually having a similar total voting power to RARI, whale domination should be meaningfully reduced.
Mitigate bribery and collusion attacks: non-transferable tokens are much harder to sell access to, so this helps, but there is still vulnerability that can’t be mitigated without true private voting schemes (e.g. MACI).
Mitigate voter apathy issues: both RARI and pRARI holders can delegate voting power. Delegation isn’t perfect (see state governments…) but it does mean average participants can “fully participate” by delegating to more active participants.
Maintain RARI importance: RARI still plays a major role here, but it does share space with pRARI (that’s the price for mitigating plutocracy and bribery).
Option 2: “R Memberships” (more similar to OP)
A new non-transferable NFT contract “R Memberships”
Anyone on a whitelist can mint themselves a new, personalized R Membership
The DAO adds users, creators, and builders who have put significant effort into the community on the whitelist
Some amount of RARI (affordable but significant, say $500 worth) can be staked “onto” R membership NFTs to activate them
Active memberships grant a high amount of voting power to the NFT owner (e.g. 10x the activation amount)
Membership voting power can be delegated in the same way that has been proposed for RARI (activating & delegating should be doable in 1 step?)
How does it do on the design goals?
Mostly the same:
Mitigate plutocracy: like pRARI, memberships would be non-transferable, go to non-financial contributors, and grant significant voting power, so this would help a lot here as long as we gave members significant voting power.
Mitigate bribery and collusion attacks: again, non-transferable tokens are harder to bribe access to, so it helps, but there will still be dangers without private voting.
Mitigate voter apathy issues: again, delegation is the solution here. Not perfect, but it should help.
Maintain RARI importance: again, RARI still plays a major role here, but it does grant less effective voting power, which is the price we pay for helping with the plutocracy / bribery problems.
Comparison & Conclusion
These designs mainly address the problem of plutocracy, or of not enough representation for contributors like users, artists, and builders. Addressing that problem, IMO, is a big deal: it would (1) help attract more users, artists, builders to the community, and these are the folks that make everything happen, and (2) probably increase the quality of decision-making, since it raises voices of “people on the ground,” so to speak.
The biggest differences between pRARI and memberships are:
Number of new people – memberships would likely be more targeted at significant contributors, while pRARI would be more like RARI, going to large swathes of people.
Role of RARI – with memberships, members have to stake a modest amount of RARI to activate their memberships, while pRARI doesn’t require anything like that.
NFTs vs ERC20 – this might be more of a superficial difference, since both would be non-transferable anyway, but there could be some cool social value to holding a membership NFT that wouldn’t be there for pRARI.
My preference is R Memberships, however with the following modification:
changed to:
The “R Memberships” NFT holders can add users, creators, and builders who have put significant effort into the community on the whitelist as a new NFT holder
I’m still processing the rest of these proposals, but I would add one additional design goal. Something along the lines of:
Avoid concentration of power: make it difficult for any single individual to control greater than X% of decision-making weight (where X is small enough that a handful of people cannot collude to gain more than 50%)
I’m pretty confident this was tacitly under consideration, but IMO helpful to make it explicit.
Also, if I may play devils advocate a bit, to what degree is it actually necessary to maintain RARI importance from a governance perspective?
100% agree on avoiding too much concentration of power.
I think there are probably a bunch of other goals that the governance system as a whole should have (e.g. to make wise decisions, to avoid gridlock). The goals I listed are meant for an experiment to get away from pure token governance, specifically. At least that’s how I was thinking of it!
WRT this:
…to what degree is it actually necessary to maintain RARI importance from a governance perspective?
It’s hard to say to what degree exactly, but here’s what I think:
Financial stakeholders are important stakeholders (they’re contributing funding, directly or indirectly!), and they should always have some representation.
Right now, financial stakeholders are a large majority of the DAO, so any changes will have to be approved by them. As a baseline, they probably won’t approve anything that feels too unfair to them!
and this:
The “R Memberships” NFT holders can add users, creators, and builders who have put significant effort into the community on the whitelist as a new NFT holder
Yeah, this seems okay to me on first glance.
Here’s a third possible experiment:
Option 3: Sub-DAOs & main DAO
Each stakeholder group gets their own sub-DAO: RARI holders, rarible protocol users, rarible protocol devs, NFT creators (we somehow decide on these)
Each sub-DAO can be controlled in a self-determined way
Each sub-DAO gets 1 vote in the main DAO (could be the multi-sig, with each sub-DAO as a signer)
Instead of Option 3 (quoted here for clarity), I’d recommend something more like the weighted approach from Dmitriy. It achieves the same goal of giving explicit voice to different groups of stakeholders, but without losing the voice of minority stakeholders within each group (which is one of the problems with the electoral college in the US).
2
Ideally, governance would be designed to optimize for the success of Rarible, however Rarible DAO defines that. I get that this token holders wanting to maintain power is the most likely tactical scenario, but I would encourage governance design to identify an ideal state free from any current conditions, and then work back from there to fit current conditions as necessary. It may be the case, for example, that token holders can rationally expect greater financial returns if they give up some of their current control, if only because governance of what they own a share of will be more robust to attacks.
3
In general I support some level of permissioning for the early part of a DAO’s life, as long as decentralization is preserved / concentration of power is avoided, so I think I like the “R Memberships” idea.
One thing to consider is how a delegate can lose their voting rights if they are behaving poorly (other than missing the basic criteria Eric outlined in the op). Can only other delegates strip them of their voting rights, or do all stakeholders (RARI holders?) have that ability?
One naive approach would be to condition a delegate’s status only on RARI delegated to them rather than their own RARI, so that other RARI holders would always have the ability to withdraw their delegation. But a couple problems with this:
It doesn’t work as a constraint for non-financial stakeholders who have been elevated to delegate status
It’s susceptible to Sybil attack by delegates, who could spread their RARI around to different accounts to and thus delegate to themselves.
I think this can be a really valuable experiment for the DAO. Currently, we have a bunch of RARI and would love to participate in the DAO more but there is honestly so much going on that we can’t keep up with it all. I love the idea of being able to have a delegate for developers that can spend the time to familiarize themselves and engage with the DAO and the proposals. I would certainly be interested in allocating our RARI to a delegate like that to engage on our behalf.
I also like the idea of splitting up the voting between RARI holders and community-nominated NFT delegates. This can help us avoid a single massive RARI holder from tanking an idea that would be beneficial to the rest of the community. It also helps create more opportunities for people to feel their voice can be counted without having the financial resources to hold a lot of RARI.
Elected delegates that have skin in the game / have more context and expertise to push things forward in the right way > pure token voting
Longer term, moving towards a model where people have more weight on decisions that they have more experience / reputation is a big improvement on just elected delegates. Curation of these delegates will be key and mechanisms to ensure no one delegate gets too much power and there’s a culture of re-electing delegates regularly to ensure the existing ones dont abuse the system to keep themselves in power.
It’s really hard to make a token truly non-transferable. Generally you simply restrict transfer for those that didn’t plan ahead. Someone looking to attack the system will probably plan ahead to make sure their tokens are transferable.
Wouldn’t this make the system more plutocratic? (or at least maintain the status quo)
If I were a whale (or anyone, really) the optimal behaviour is to obtain an R Membership and stake my balance on it, thus 10xing my vote weight. If I can’t stake my entire balance into my own membershipt, the optimal behaviour is to bribe the other members and delegate my vote weight to them.
It’s really hard to make a token truly non-transferable. Generally you simply restrict transfer for those that didn’t plan ahead. Someone looking to attack the system will probably plan ahead to make sure their tokens are transferable.
Can you explain? I agree that “true” non-transferability is basically impossible, but DAOstack DAOs and Moloch DAOs have been using non-transferable tokens pretty effectively to govern high value treasuries for a while, no?
The DAO should give these tokens to addresses it thinks will not want to betray the community’s trust – addresses that have built up social reputation as contributors, for whom getting caught doing a tricky account custody sale or whatever would be very costly (and wouldn’t really be worth that much on the market, as a single membership couldn’t ever be a huge voting power whale).
Wouldn’t this make the system more plutocratic? (or at least maintain the status quo)
If I were a whale (or anyone, really) the optimal behaviour is to obtain an R Membership and stake my balance on it, thus 10xing my vote weight. If I can’t stake my entire balance into my own membershipt, the optimal behaviour is to bribe the other members and delegate my vote weight to them.
Let me clarify: staking on memberships would be capped, and the cap would specifically be affordable (say, $500 worth of RARI). The idea is to ensure members have some financial stake without favoring whales.
Great topic, @Rarible@m0zrat@spengrah a lot of truly interesting thoughts.
Overall thoughs: I see a demand for more power for users/builders/other groups of people who’d made a valuable contribution to the ecosystem. And some proposed changes of the governance model on how to mitigate this.
My overall perception is that the system with two tokens is complex, there is a 50/50 balance of power.
The biggest questions to date are:
How do we maintain an alignment of RARI holders vs contributors in the common goal.
What is the economic incentive for contributors to thrive protocol adoption?
How do we maintain the simplicity of governance so that new people joining can understand that reading 1 paragraph of text?
There is some sense in including a reputation model. My thinking is that it should be included not in a “Percentage of Votes” model, but as an overlay like veto power, etc. One more thought is that sheer publicity of some of the key voting delegators would add a reputational component in the equation. An example of that would be a16z delegating to a reputable entity like Universities blockchain society and research groups
When we designed the system at the first place there was an intended distinction of havng more than Half of the total supply to be distributed among the community. Maybe we can design a system where contributors just have more RARI and everyone is aligned with a single economic incentive?
Definitely. I would love to find simpler options! That said, I think we could make any of these simple enough to be workable, and some kind of multi token or multi-part model might be necessary to have a strong DAO long-term.
I think alignment is fairly strong in these ideas. Let’s take the “memberships” option as an example. Both tokens, RARI and memberships, grant some voting power in the DAO, so holding the tokens creates basically the same incentives for holders: to have the protocol & DAO thrive. The difference would mainly be in the actions you take to acquire those tokens (e.g. purchase or earn RARI vs. only earn a membership through actions).
Curious to hear more of what you mean here! There are many directions the designs here could take so it’s good to throw as many ideas at the wall as we can at this stage.
The tricky thing here is that we are seeing many of the people who earn RARI via using rarible.com simply sell it immediately, right? Even if small holders do hold their RARI, there will be more and more ways for them to rent out their voting rights to the highest bidder. Since small holders have relatively insignificant voting power, they may prefer that. RARI liquidity mining is also designed to slow down over time, right?
It seems overwhelmingly likely that, overtime, RARI will be more and more concentrated, unless we change the design somehow.
We might give out RARI to a creator b/c we want creators to be represented, but if they sell it or rent it, that RARI is no longer representing a creator perspective, it’s representing an investor perspective. These experiments all have some element of non-transferability because non-transferability seems like a way to maintain both contributor and investor perspectives – I believe a DAO that makes sure creators have a voice will be more successful for everyone in the long run.
My thinking is that this can be pretty easily mitigated with social consensus rules
for example:
maybe on the NFT Membership side, it is 1 token 1 vote
maybe you have very strict requirements to stay a member (for example: you need to attend calls, and participate in governance: ie, active members)
maybe there are a set of rules or principles that people need to adhere to in terms of staying a member
But also, I’m not sure how it’s opaque: it’s very transparent who these people are. It should be people who are most active in the community, and this is fairly easy to measure / track / ensure.
I should add, I do think it’s really important to experiment with this stuff. I really dig all of your suggestions, @Rarible, @m0zrat, and Spencer.
Limiting plutocracy is incredibly important but also a huge challenge.
Along with the suggestions above, I’d be really keen to explore straight up quadratic voting as an option. (must acknowledge the obvious Sybil resistance requirements)
