A Roadmap for YAM (Part 2 - Governance)

Part 2: New Governance Models and Upgrades

If you haven’t read Part 1 on re-focusing vision, please read it first.


In the 6 months that I have spend contributing full time for YAM, I have experienced first hand the hurdles, problems, and opportunities that all distributed organizations like ours face. While there are many elements of the DAO that still are being worked on and figured out (who we are and what we do, finding PMF with our products, etc, etc.), the one thing that underpins all of this is our ability to govern YAM. While we have made great strides in moving DAO governance forward and I am proud of our commitment to on-chain decentralized governance, there are still pain points that we need to solve:

  • Lack of governance participation in both voting and active DAO participation
  • Lack of incentive alignment to reward long term participants over short term participants.
  • Reliance on outside users to provide liquidity for the YAM token, with an indeterminate amount of future issuance required to maintain it.
  • Reliance on key members of the community to implement governance actions in code. YAM is overly reliant on these individuals and this leads to over-taxing them and creating added stress for them.
  • Our governance process relies heavily on on-chain transactions and on-chain voting that is very expensive and requires bundling many transactions into large monthly votes to be economical. This adds overhead and stress for Devs who need to make sure everything is included and accurate while battling artificial time pressures.
  • In the current model, governance is either decentralized and expensive, with lots of overhead, or centralized and easy. We should look for ways to find a middle ground between these 2 extremes so that the DAO can be run efficiently, but also with protections and decentralization built in.

The following are my thoughts on the state of governance and a call to action to start researching solutions to solve them.

Governance Participation

Why so little?

This is not a unique problem among DAOs in the web3 space and there are a number of different models that have been created to try and solve it. The first question that must be asked is “why is there so little participation?”

  1. Lack of time
  2. Lack of incentive
  3. Lack of agency (perceived or real)

Lack of time is a pretty straight forward reason. Many DAO token holders are busy people and did not acquire tokens with the intent to spend time participating in governance. They often view their tokens as investments, and like typical stock holders, have little interest in actively participating in the governance of the DAO.

Lack of Incentive is a related issue to lack of time and comes in 2 forms. The first form is the assumption that a token holder doesn’t view their contribution as necessary and that the DAO will do just fine without their participation. They do not believe that their investment is improved by their participation. The second form is more related to the lack of time, and is simply that it often does not make economic sense to spend the time necessary to participate if that time is not compensated directly. Investing in something that requires active participation is a very different beast from investing in something that can be held passively.

Finally, Lack of Agency is a combination of both the above reasons in which users do not feel that their contribution is appreciated, or do not know how to contribute meaningfully. Token holders who feel agency and control over some aspect of the DAO are more likely to participate productively. Requiring users to vote on things that they neither have the time or expertise for is counterproductive.

Finding ways to mitigate these issues is central to improving participation.

How to fix it? Process, Delegation, and Incentives

Process: Getting things done while remaining decentralized

Pessimistic Governance (inefficient)
The way YAM currently works requires that any time the DAO treasury is interacted with, a quorum of token holders needs to agree. Althought YAM pays full time contributors to do the day to day work to make the DAO run, those contributors have no power over the treasury other than via their own token holdings. The current model works like this because this is how our governance process checks the power of the contributors and ensures a certain level of decentralization. This model optimizes for maximal security for the treasury and trades off efficiency.

Optimistic, centralized Governance (insecure)
Another other model that many DAOs use is to hold funds in a multi-sig controlled by a few key contributors, who act on behalf of the tokenholders and are bound to follow their wishes by social contract. This model relies on trusting the signers of the multi-sig to both be honest and to execute the will of the token holders. This model optimizes for efficiency and trades off security and empowerment for token holders.

Optimistic, Decentralized Governance (In between)
This is not really one model, but a spectrum that lives between the extremes of Pessimistic and Optimistic governance. It utilizes both models simultaneously to check the powers of contributors, while allowing them more freedom and efficiency, with a fallback to pessimistic governance if needed. This is where I believe we need to be focusing our efforts to improve governance and it has the widest design space to customize to any particular DAO’s needs.

An implementation of an optimistic, decentralize governance process could work as follows: A core group of 6 contributors is empowered to do treasury management and are given control over a specific subset of treasury funds with a 6/6 multi-sig. They work under a specific mandate and make their work public so that the larger community of token holders can follow along. Any transactions that they perform are timelocked for a period long enough where token holders can veto the action via full governance votes. If no one objects to the transactions that this group performs, there is no need for a vote.

This leads to fewer votes, and the votes that do happen will tend to be more important.

Delegation: Levels of commitment

Some token holders want to participate in governance and others don’t. Giving token holders the choice to delegate tokens to others who are available and interested in diving into the details needed to make decisions is an important feature. Delegation is an opportunity to organically provide incentives to delegates who take the time and effort to govern. If governance is rewarded, then delegates can earn for their work without needing to be explicitly hired to do day to day work.

YAM currently has delegation, but the interface to use it is difficult and gas costs make it expensive. We also don’t have any infrastructure built around coordinating with delegates.

Incentives: Governance is Work

If governance is an important part of keeping a decentralized project or DAO running, then it is inevitable that it will take work to make it happen. Some governance is performed by contributors who are paid to do day to day work. But if we want our organizations to truly be steered by a community of token holders then they need to be compensated for their time (and gas costs)

Beyond simply being compensated, DAOs must consider how they can align the incentives of their contributors, token holders, and customers to govern effectively. Simply paying token holders to vote doesn’t ensure that they will govern effectively or in the long term interests of the DAO. With YAM being available to purchase on AMMs and CEXes, aligning incentives also protects against governance attacks.

Long term incentivization can be done in many ways:

  • requiring locking tokens to vote
  • giving more voting power to longer locks
  • rewarding governance participants based on future KPIs


There are multiple new governance models that are being developed that approach these problems in different ways. We need to understand what the unique needs of YAM are in order to come to a conclusion about what platform to use. Below is a non-exhaustive list of different tools, implementations and options for extending governance

Zodiac (previously Safesnap)

A series of tools built around the Gnosis Safe multi-sig wallet intended to allow for it to be used with greater flexibility by DAOs. These tools add the ability for snapshot to be used to execute on chain actions, cross-chain communication, guardrails, and more.


One that has been gaining prominence recently is the veCRV model that Curve uses for their governance and gauge voting. Sushiswap has discussed using this model, it has come up in the new tokenomics discussions at Yearn, and on a smaller scale, PieDAO has implemented a version for themselves that just went live: Own the DAO - Walkthrough of veDOUGH | by Alessio Delmonti | PieDAO | Medium.


Colony.io is a new governance product built on xDAI that incorporates many features to run a DAO out of the box, with some novel features like reputation and optimistic voting. This is a full suite of tools for governance.


DxVote is the upgraded governance system being built out by the DXDao for their own voting purposes. It merges their 2 token model into a combined voting scheme that requires both DXD and the non-transferable REP token to vote, along with voting incentives and other features. old post that goes over the high level ideas. Governance 2.0 - Signal Proposal - dxDAO - DAOtalk


Another option (or additional layer) is a system like SourceCred where contributions to the discord, forums, and github are rewarded via a social graph that is created specifically by YAM to focus on the things we think are important. This is an “automatic” mechanism where specific actions earn cred (and subsequently grain) which can be redeemed for YAM, or a portion of revenues, or whatever else we choose.


Coordinape is a tool used by teams to reward contributions, where the reward distribution is more manual than something like SourceCred. The social graph and subsequent contribution payments are distributed directly by other contributors. This would work well for situations where contributors are known and are working together. The core contributors to YAM will soon be doing an experiment with Coordinape.


It is clear to me that we need to spend the time to make our governance process work for contributors and users alike. That starts by understanding what works and doesn’t in our current process, as well as how we want a new governance model to work. This is a stab at starting that conversation. This post in not exhaustive and could probably use some editing, but even incomplete, this is something we need to be thinking hard about.

We started out as an experimental on-chain organization. We have had successes and failures, but to me it is clear that the model we are using is naive in its implementation. In the spirit of that initial jump into the unknown, we should evaluate where we have been so that we can look toward the future and continue to innovate with our DAO’s governance process.

Further reading

  1. Building and Running a DAO: Why Governance Matters | Future
  2. Decentralized Governance Structures | by Joey Santoro | Fei Protocol | Medium
  3. Andre Cronje’s oSushi post: Sushinomics : introducing oSushi - Proposals - SushiSwap
  4. Long and extensive thread from the Maker Forum about paying governance delegates: Delegate Compensation [informal poll] - Governance - The Maker Forum

Part 3 covers my thoughts on Tokenomics and token issuance.


Quick follow up here on Governance. There have been a few developments in the space that are worth taking a look at. The vote escrow model pioneered by Curve has grown further in popularity in the space, with multiple large projects looking to use a version of it for their needs.

Balancer using Vote Escrow

Fernando from Balancer just yesterday posted a proposal to revamp their governance and tokenomics models to use a vote escrowed model. You can read the whole proposal here.

While there are lots of details to dig into, one of the most interesting to me is their use of 80/20 LP tokens as the core governance tokens to be locked. Of course this makes perfect sense for them as they are using their own protocol, but it is also something that I have considered as smart for any governance system. I wrote a post on protocol owned liquidity, which goes into ways for the DAO to own liquidity and assure that anyone can join the DAO without eternal incentivization. I believe that some amount of locked protocol owned liquidity is necessary for the DAO, but determining how much is a hard question.

How would we use this?

Using liquidity pool tokens as governance tokens helps answer this question because it creates a lower bound for the amount of liquidity that we can expect based on certain assumptions about how many tokens are locked in governance.

If we can get 50% of the token supply locked into governance to earn voting rights and revenue rights, that would be approximately 7.5M YAM. If those YAM are first deposited into an 80/20 Balancer pool then this would be equivalent to about 1.875M of liquidity (7.5M / 4), which is right around what we currently have in the current sushiswap pool.

Getting 50% of the circulating supply locked into governance is definitely ambitious, and these numbers may need to be adjusted. But the point stands that we could achieve similar results to what we are getting now while also incentivizing governance. The nice feature of this system is that even if participants who lock their liquidity in governance don’t vote they are still providing liquidity.

If we implement the tokenomics model discussed in this post, at current prices (~$0.30/yam) that would be $2.25M worth of YAM locked and a treasury size of ~5.5M. Lets assume 1M held aside for operations and that a 2x multiplier on yield earned by the treasury. Even if those funds were conservatively deposited into Yearn yUSDC/yDAI and crvSTETH/crvSETH pools we would be able to earn governance voters twice what they could yield on their own (5-7%) while also providing liquidity.

Protocol owned liquidity Success tokens could still be used to further increase liquidity and provide a baseline minimum. Once we get a sense for how much liquidity the governance locking process would provide, we can craft the success token initiative to help us reach a point where we believe it is safe for new initiatives like the YAM fuse pools.

MakerDAO and strong Counterarguments to the veCRV model

Here is a new proposal on the MakerDAO forums proposing moving to a vote escrow style model. [Signal Request] gbMKR, a proposal to benefit long-term MKR holders - Signal Requests - The Maker Forum

The whole thread is interesting with lots of good points, some of which are specific to Maker and others more general. Worth reading if you are interested in different views about the veCRV style tokenomics.

Monet Supply’s post criticizing the CRV model in general is very well written and makes a lot of strong points that bear consideration, especially for projects like YAM which have different governance needs than Curve. [Signal Request] gbMKR, a proposal to benefit long-term MKR holders - #16 by monet-supply - Signal Requests - The Maker Forum

I don’t yet have a conclusion from these posts, but I think there are pieces of these models that we can use to effectively govern YAM to the benefit of all holders and community members. I will add more thoughts here as I keep thinking about these issues.