RFC: DAOs Committees: Balancing Decentralization with Efficiency

The following is a post that brings a discussion to the forum that, given the current state of Arbitrum’s governance, I believe would be beneficial to have. It concerns the use of committees for resource management. This is just a general outline of the issue, and I would like to receive feedback from the delegates to know if there is interest in delving deeper into the matters presented here.

The article published by tokenbrice on their reasons for resigning from the GHO Liquidity Committee caused quite a stir. I encourage every participant of DAO governance to take time to read the article, since it raises many valid points that occur and are replicated across many of the most prominent DAOs.

Lately, Arbitrum DAO has opted for the formation of multiple committees to manage various aspects of resource allocation decisions and the creation of frameworks.

The establishment of such committees has been the response to avoid “centralization” of decision-making regarding resource allocation in the hands of a few or a single person. Whether this is the best approach is debatable, as is the advisability of centralizing certain types of decisions (an interesting discussion about this in Vitalik’s article on convex vs. concave decisions).

I don’t think it’s wrong for a DAO to experiment with appointing committees to sort things out. However, I’m Argentine, and in our country, a phrase from an old political leader is very popular, stating that “If you want something done, name someone responsible. If you want something to be delayed forever, appoint a commission,” (this phrase is originally attributed to Napoleon and before him to Joan of Arc). Obviously, this view stems from recognizing the efficiency of centralizing decision-making versus the inefficiency of delegating it.

In environments where avoiding centralization is sought, delegating decision-making to multiple parties seems desirable. However, forming committees without taking certain precautions can lead to an outcome as bad or worse than centralization: complete inaction. It can make the process non-transparent, decisions taken for unknown reasons, lack of accountability, or failure to meet objectives without understanding why.

Arbitrum has recently approved 4 committees (under different names):

Treasury Screening Committee

LTIPP Council & Advisors

ArbitrumDAO Procurement Committee

Arbitrum Research & Development Collective

Note that I am not pointing out one or another committee in particular nor do I claim that this should affect its functioning. The experience collected in the aforementioned committees should help strengthen what can be a framework in the future.

If we can agree on certain guidelines or best practices for the operation of future designated committees (or even apply best practices to the mentioned ones), the shortcomings of “decentralizing” decision-making will be reduced, while we can create a framework for the future that is included in the upcoming proposals.

I’ll leave here some ideas to start with, hoping to receive everyone’s feedback.

Problem 1: Lack of clear tasks and objectives to measure outcomes

It is crucial to define the scope of work for the committee, as well as to clearly set out its objectives in order to later measure the results obtained. Just as a protocol requires metrics, KPIs, and milestones to be granted funding, a committee should be required to have clear and defined objectives with deadlines to achieve them.

Problem 2: Lack of external accountability

Committees should have a designated leader responsible for ensuring that milestones or deadlines are met. It often happens that when everyone is responsible, no one is responsible. If there is a concern about centralization, the leadership role can rotate. However, there should always be someone accountable for the status of the work and the adherence to deadlines.

Problem 3: Lack of internal accountability

The roles within the committee must be clearly defined. Each member should have clear objectives and deadlines. It’s important to avoid having a single, generic goal for the committee, which can cause some or many members to feel lost about how to act or coordinate.

Problem 4: Lack of transparency

Communication among committee members should be as open and transparent as possible. Decision-making should also be as open and transparent as possible. A Zoom, Google Meet call or telegram chat are neither open nor transparent, even if recordings or transcripts are shared later (which I doubt many will watch, to be honest).

As uncomfortable as it may be, an effort must be made to avoid repeating the mistakes of traditional politics and to overemphasize transparency and communication. There are tools available for holding meetings and avoiding spam, such as Discord (where open channels can be created for users to read/listen without interacting).

Someone can be designated to handle reporting or to summarize and communicate all meetings.
Communication should always be a top priority.

Problem 5: Lack of engagement with governance / community

Specific mechanisms for feedback from the broader community to the committees and vice versa could be further elaborated. This would enhance the iterative learning and improvement of the committee’s operations.

Problem 6: Are committees a popularity contest?

The effectiveness of a committee does not only depend on having clear objectives, accountability, and transparency but also on the competence and confidence of its members to carry out their tasks.

Is an open election the best way to choose committee members? How can we ensure that its members are those better skilled for the task? (this is an open question tbh).

What’s the best mechanism for assessing the skills and qualifications of potential committee members?

Is it possible to avoid popularity contests at all? One solution could be providing a budget for training committee members. Proper training can equip committee members with the necessary skills for informed decision-making. This could be a useful tool in this context, where committee members come from a wide range of backgrounds and may not have previous experience in similar roles.

Open questions that would love to get feedback on:

Conflicts are inevitable in any group decision-making process. Which is the best way to handle conflicts within committees?

How can we manage conflicts of interest between the candidates for a committee and at the same time ensure that it is made up of the best qualified people? Is there a way to minimize the risk of conflict of interest without harming Arbitrum? What is the balance?

Conclusion

This post is not intended to be a framework, but rather to be an RFC that seeks the engagement of the DAO to understand if it makes sense to create provisions of this type that delegates should expect or require when the formation of committees is proposed in AIPs.

I also plan to create a version 2 of this overview, incorporating examples from this and other DAOs to illustrate what has been effective and what hasn’t. This will help clarify which frameworks are most likely to succeed. It would be great if current committee members share their own opinions and insights on their roles.

I await your feedback on all the points raised here :slight_smile:

7 Likes

Interesting post.

I have found myself reflecting on this a bit. I think that every community/social entity, growing up, would go naturally through the following phases accordingly to the scale of numbers

  • initial set of founders/leaders, all with specific skillset, all deciding together
  • still founder/leaders, but then start involving more community members, especially the one actives or with skills
  • then, full out democracy on 1:1 based decisions with people being appointed by each others/leaders
  • finally, group of people being appointed for specific functions as the scale of work is too much to handle for a single guy.

Now, this doesn’t refer to any dao or our dao. But to any org. Is the less friction path. For which I don’t see necessarily anything wrong, as long as there is no stigma of a group taking a lead versus a single and viceversa.

On your questions:

  1. Agree. Every committee/appointed person need tasks to be clear. Problem: the proposals, that go first on snapshot then on tally, have more often than not holes, and is sometimes impossible to establish a scope. Crafting the proposal in a more formal and robust way on the other hand would take a lot of time.
  2. Would be good. Only doable if the leader has power on the rest of the committee. This, usually, translate in the ability to alt/change the compensation for example. Do we want to go down that route? Or, is there any alternative mechanism?
  3. I think is the same problem of 2.
  4. We can for sure getting better with reporting. But is not only the report: is also how is written, where is published and the language used. A standardization would help us a lot. Difficult to achieve in a corporate world, VERY difficult in a dao
  5. this is a broader problem. Again, I agree, we need to use the forum more, and we need people to also write less (me included kek)
  6. sometimes yes. The idea of training members is interesting tho, making people committed to the dao grow, inside the dao, would be something matching people in corporate being paid to study to bring after more knowledge and value.

On conflicts: the way crypto works, is quite impossible. Anybody can have affiliations to a lot of entities, either directly or indirectly. Full disclosure, and alternative in numbers when is time to recuse from a specific vote (so a pm can take the role for example) is key. But only using intellectual honesty will take us up to a certain point, and bad actors are always ready to exploit this. Yes, we need to think this through.


I have my own idea tho.
People who work for the dao (cause this is the term we need to start to use. If the dao compensate you to do something, you are working), should be threated most of the time like managers, not employee.
This means a compensation structured in a way that a base salary is there. But then, the bulk of it comes if results are effectively met. This poses the followings:

  • who evaluates that the results were met? (probably the dao)
  • how is the evaluation done? (probably through a vote)
  • what are these “results”? (probably tied to 1., re: clear tasks/goal)
  • is it ok, for a “manager” to be evaluated by people that might or might not have that same level of knowledge on the topic? (answer should be yes if i am in a dao, then again, if i realize a risk model that saves in the course of 2 years 5.6M dollars in costs, what instruments have effectively voters to evaluate this?)
  • how can we keep the mechanism above, without disincentivizing people to work for the dao? (there could be people who effectively need money upfront, or that might feel not up to the task. Answer could be: optionality of payment).

I know all of the above rotates around compensantion. This is not cause I am a greedy motherfucker*. But because, effectively, in the situation in which we are, with how easy mobility is for a single person through several DAOs compared to normal job, the only real way to align incentives and success is through the token itself which is an expression of the whole DAO.

*(well I probably am but that is not the point here)

3 Likes

Hello, @pedrob
I like the topic you brought up.

  1. I agree with you about the clear goals of the committees, which I previously expressed to the authors of proposals to create these committees.
  2. I also like the idea of appointing a responsible committee member who must present the results of the work of the entire committee (of course, with a large salary for this)
  3. As for conflicts within the committee, perhaps it should be like in a jury trial - until everyone agrees with the decision, no decision should be made. And dissolve the committee (re-elect) if such conflicts do not allow this committee to work.
  4. As for the qualifications of members, this is a difficult question, since for the most part there are no diplomas in the crypto space that confirm your qualifications and everything is based only on your experience and reputation.
1 Like

Thank you for raising this topic. I would like to share my thoughts on some aspects of the issue.

It is essential for the community to understand the inherent corruptibility of the committee mechanism.

These phenomena are very familiar to the Chinese people. China’s governance mechanism is ostensibly a parliamentary system, but in reality, it is a one-party dictatorship committee system. The power of the entire country is decentralized to one committee after another.

It operates on the basis of compromise, an unspoken mechanism. Each member expects that his concession on one issue will be met with concessions from others on other issues - the space of internal opacity allows committee members to compromise with each other and compensate each other in the future, thus quickly finding common ground and promoting action.

The community has a false imagination about the source of the efficiency of the committee: experts form a committee, and based on logical and factual debate, form rational judgments to quickly advance the cause. However, this is a beautiful fantasy of intellectuals. In reality, the main reason why the committee mechanism is more efficient than the parliamentary system (proposal-voting) is because because experts can compromise with each other and reach a consensus more quickly. The efficiency of the committee mechanism relies on internal opacity and模糊全责. It is more suitable for political actors who are good at compromise and waiting, while researchers (such as Tokenbrice) always find it difficult to survive in committees. Questions 3, 4, and 5 are actually all derived characteristics of the committee mechanism.

In short, the committee mechanism is a trade-off of efficiency for governance with impurity and public corruption.

1 Like

Thanks for kickstarting such an important discussion! As a creator of one of the aforementioned committees (treasury screening committee for STEP), here are some insights i’ve gained

This is perhaps the most important one - to recruit high quality committee members, we need bounded tasks without too much scope creep.

The way we recruited members was laying out 3 tasks they need to undertake: draft an RFP post snapshot vote and pre-tally vote; screen out applicants after submission window closes; draft/approve report cards on shortlisted applicants for delegates to read before the snapshot weighted vote determining exact allocation between them

This is made much easier when committees have dissolution baked in from the start.

In general we need to differentiate between posts (regular elections to hold the role, like ARDC/procurement/ltipp council will become if deemed successful) and committees (timebound and dissolved after achieving a purpose, like our screening committee which is dissolved after allocations are complete)

The MOST important thing here is do not give committees direct power to award business. That is and should remain the job of delegates and ARB token holders, as jojo puts it

For STEP, we found the tokenized tbill RWA space is quite small and there was an existing network of favors, relations, etc between players in the space. So rather than give power to one set of players (committee members) to award huge business contracts to another set of players that they may work for in the future or take other favors from, we circumscribed their role to a yes / no vote for consideration by the DAO and nonbinding opinions which delegates can read on each provider before voting

Its good practice to have a nonvoting member serving as facilitator (who only votes in case of a tie between other committee members). Usually, the best person to play this role is the creator of the framework or the one who brought the committee together, so they can also do the engagement with governance (as they did it earlier for passing the proposal in the first place). For example, I serve the facilitation role in the STEP screening committee.

Compensation for this facilitator should be the same as other committee members.

This is perhaps the trickiest part of it all. We should remember these are prestige postings and so not overcompensate; but also pay an amount that you can demand their time and don’t have to beg for it.

I’ve also found it helpful to peg their payment not to some arbitrary number , but to an external parameter no one really has control over. For example, in STEP we pay 500 ARB per member per valid application received.

1 Like

I wholeheartedly agree that traditional organization hierarchies work; modern businesses have had decades to refine to something that works today, however, whenever we try to impose this type of structure, it is met with vehement dispute, and I believe this is due to a mismatch on trust vs trustlessness, and trust is needed for this type of model to work, and we live in a don’t trust, verify world. Accountability and defined responsibility are key as a starting point, and being held accountable by the dao, I think, will never scale. So we are in a weird limbo right now where trust is being established; ultimately, we will probably need something like a tri-party system executors/workhorses to get shit done, a board of delegates to hold the executors accountable and course correct them when needed “mgmt”, and some police/judicial group to monitor and investigate any contention or conflicts and bring that evidence to the dao to vote/enforce.

But as it stands today, we are not ready to trust anyone with any significant amount of capital and to have any kind of unilateral control. I think eventually we will move out of this growing maturity phase, but until then, the only things that are passing are committees and pilots, which imo is better than outright deadlock.

2 Likes