Modular Governance

Published in May 2022.

Governance can be described as the art of wielding power, taking on responsibilities, and making decisions over a group of coordinated individuals in order to streamline efficiency, reduce confusion, and resolve conflicts. Decentralization holds great promise in giving everyone a voice, as opposed to concentrating power in a single entity or individual. However, I believe that centralization has been unjustly criticized and its benefits have been obscured. The benefits of centralizing power are evident: it allows for a faster pace and a more efficient system as responsibilities are clearly defined within each layer of authority. This is why companies like Amazon, Google, and Facebook have created tremendous value in society and generated millions of jobs using a top-down approach. The current DAO structure works well for executing public goods or non-profit projects where community engagement is of top priority. The question remains, can DAOs evolve and achieve the efficiency required to be a legitimate alternative to the traditional corporate structure? Will future startups be built as DAOs? This essay explores the modifications necessary to make this leap.

Whenever governance is discussed, the voting mechanism often takes center stage. There are various types of governance voting: first, there’s the one-person one-vote model, where everyone gets one vote regardless of how much stake they have, how wealthy they are, or their rank in the hierarchy; this is exemplified by presidential elections and enabled by the use of Soulbound Tokens in web3. Next, there’s the one-token one-vote model, where the person with the most stake in an organization will have the greatest voice in decision-making. Lastly, there’s the custom model, where voting power can be manually configured for a select group of individuals and be based on attestations such as reputation tokens or other scoring systems. There is no superior voting structure; instead, the most effective DAOs will adopt a custom model, incorporating several governance structures and manually adjusting based on their own unique structures. It is crucial, therefore, to provide configurability and flexibility for DAOs to create a governance structure that fits their needs.

Reputation Token

The challenge: the size of token-holders’ stake in a DAO often doesn’t correlate with their expertise or understanding of the problems that proposals are trying to address. This leads to an unfair allocation of voice and misinformed decisions in crucial proposal calls. DAOs need to shift away from fully basing voting distributions on token-holders and favor reputation-token holders more. A reputation token is a fairly new concept in the DAO tooling world, but it basically exists to help gauge contributors’ value in regards to a particular field by assigning tokens proportional to their contribution (quality and quantity). There are many projects working on this domain, such as SourceCred, Coordinape (value mapping), and DegenScore (web3 reputation).

SourceCred uses the classic computer science concept of graphs (nodes and edges) to illustrate how value flows inside an organization. Nodes, represented by contributors, can send contribution points to other nodes and mint points from the “seed” by performing certain actions. An example of the former case is, Person B gets 10 reputation tokens for every like he gets from his post, and the latter would look like, Person B gets rewarded 20 reputation points for every post he makes. Again, the cases and conditions are configurable through the open-source nature of the primitive’s code. Another possible way to distribute other than basing it on actions taken is rewarding contributors based on how long they stay in the DAO.

The issue with SourceCred and DegenScore is they fully measure reputation based on quantitative metrics, which is a very poor approach in most cases. The number of votes you pushed or how active you are on Discord may provide an indication of your involvement in the community, but doesn’t necessarily measure the value you contributed. Coordinape is an awesome tool that gives each member of the community 100 GIVE tokens to be allocated to others at the end of every working period (known as epochs), but this approach has a high degree of inaccuracy in rewarding optics of working and may be subject to bad actors. A contributor may receive high social capital by being very active on Discord, but in reality, he’s just spamming memes and meaningless content to channels and contributing nothing in community calls.

In short, the reputation system shouldn’t aim for full decentralization nor a quantitative approach. We lack a qualitative method of measuring contributors’ worth, and the only way to achieve this is through federation and leadership in DAOs. For example, in Superteam’s governance, you can’t reward contributors solely on the basis of project or task completions because not every project weighs the same, and it would make it unfair for contributors who do more complex but fewer tasks. A federated approach to this is to allow product managers to manually set XP points to tasks, agreed upon by the project members. Since this is goal-oriented, an XP multiplier will come into play if the project exceeds its pre-determined goals. The downside to this is that, since the process of XP allocation is very subjective, it will make it hard for people to understand the score.

I don’t believe one approach is better than the others, as I am also against the one-size-fits-all approach to this. In my opinion, there will be use cases where quantitative methods are used more extensively than qualitative approaches, and vice versa. However, in Service and Operational DAOs, I believe that quantitative measurements will be built on top of qualitative ones. For example, for every proposal you make, you will earn 3XP points, or for every mirror post, you’ll earn only 5XP, but every comment and like will add an extra 10XP on top. For the mirror example, contributors will have to manually submit the articles to be reviewed. This can also be in the form of a range of acceptability; for example, 10–20 posts a week is deemed good with no penalty, but if it drops to 1–4 posts a week, you will get penalized 10XP points. This modular and composable approach introduces another problem where every DAO will have its unique reputation structure. A possible solution to this is to encourage DAOs to create their reputation manuals so that outside people can understand what these scores mean.

Lastly, it is important for these reputation points to be on-chain or standardized off-chain, for reasons that will be explained below and in the remaining parts of this blog (mainly for modular governance; you’ll see the connection).

The way I see it transpiring is that small DAOs will have a single reputation token, and large DAOs will have multiple subDAOs (or pods) that can be functional or project-based, and contributors will be assigned to those working groups and earn reputation tokens unique to the pods. These unique reputation tokens will be useful for creating custom voting distributions for proposals that come from different areas. For example, let’s say one of DAO A’s pods is the Marketing Pod. The voting distribution of all proposals that affect the general workings of a DAO, i.e., the token supply, will allocate 90% of the distribution for token-holders and 10% for DAO reputation token holders, mainly because a minor modification in DAO A’s tokenomics can influence investors in a big way. On the other hand, proposals from the Marketing Pod will have 50% of the voting distribution for the Marketing Pod’s reputation token holders, 20% for the DAO reputation token holders, and the remaining for the DAO token-holders because decision-making in pods needs to come from people who have skin in the game.

What’s the difference between reputation-token holders and pod-specific reputation-token holders? A valid question, as this touches on my next point: all of these are programmable and composable, meaning anyone can build various applications or iterations on top of the primitive, in this case, it’s the reputation token. In the example above, the general DAO reputation-token holders is a possible iteration I had in mind of equally splitting the voting distribution pie among the pods’ reputation token holders in DAO A.

Note: Reputation Token = Non-Value Inflationary Token

Breaking down a DAO into manageable pieces or pods helps in maximizing efficiency. But fractionation isn’t enough to reach optimal productivity; both pods and the general DAO need to have clear roles and responsibilities for all of their contributors. I understand that traditionally, tall hierarchical organizations have the main problem of a long decision-making process and leaving the employees with no say whatsoever, but I believe there has been an unfair negative perception towards this structure, as people just focus on the downsides without understanding the true value proposition of it. In my opinion, hierarchical structures are a means to set clear roles and responsibilities for contributors, where these levels are used to show the level of superiority in decision-making that helps with efficiency. A flat structure with no clear responsibilities will move slower than a hierarchical structure or give rise to other problems like underutilized working hours, confusion, or severe burnouts. Thus, it’s important to point out that efficiency comes from clear responsibilities, not the type of structure. Hierarchical structure doesn’t equate to bureaucratic leadership; it’s just a way to show clear responsibilities. The sweet spot is when power isn’t centralized on a single entity/member but still has a clear distribution of responsibilities.

An example of clear roles and responsibilities would look like having a position called the Marketing Commanders, responsible for selecting projects, assembling teams, and managing the teams’ progress. Every role has different responsibilities and permissions, and all of this can be easily custom-configured by the DAO. An example of having different permissions is a Marketing Treasurer role within the Marketing Pod that can issue new money transfers from the treasury after a proposal has been approved (although proposals and execution can be automated via smart contracts, discussed later in the “Trustware” section). Furthermore, DAO A can limit proposal creation for contributors who are at a minimum of a Marketing Padawan role, to prevent low-level proposals or proposal spamming. In addition, every role also has different compensation levels and rewards in the form of reputation tokens. At last, roles are fundamentally a status symbol that incentivizes contributors to stay in a DAO so that they can level up to earn higher-ranked roles to enrich their resumes.

Avoid Voting Where Possible

This saying accurately depicts the problem in DAOs right now, where all contributors are encouraged to vote on all proposals coming from many different places. Voting needs to come from contributors who have earned their voice through their active contributions or capital investment.

In fact, this is why Socrates was against democracy. It wasn’t that he supported the authoritarian model where decisions are solely made by the elitists, but rather he was against putting much weight on voters that have low rationality and understanding about what they’re voting for. Voting should only be open for people who have skin in the game or a deep understanding of the current political landscape, and only this way can we create a prosperous society. The current model of democracy by birthright gives rise to demagoguery where leaders are elected by arousing civilians against the elites, and this often results in the stalling of economic progress. We should transition to an intellectual democracy model where a vote is earned by contributing and having skin in the game, not given. DAOs should have the ability to create constraints, for example: proposal voting is limited to contributors with a specific badge that indicates skin in the game, or contributors that have at least 100 reputation points (even though putting a constraint on reputation token holders isn’t needed because they are proportional to contributors’ value or skin in the game). These constraints are made possible by adopting on-chain attestations.

Besides that, voting is just a tedious process that can overwhelm contributors because it is essentially an extra responsibility that requires understanding and judgment calls. Proposals can be categorized further, and every proposal category can have its own voting structure and constraints. For example, proposals in the investment category have a voting distribution that looks like this:

• 50% DAO Token Holders (One Token, One Vote)
• 20% Reputation Token Holders (One Token, One Vote)
• 20% Finance Pod Members (One Person, One Vote)
• 10% OG Members (One Person, One Vote)

AP Actions

Not every decision should go through public proposals because it requires many contributors to vote, which is something we don’t want and is clearly unnecessary in most cases. Introducing Assumed Passed Actions — or AP Actions, in short — is meant to bring the efficiency that centralized organizations offer by giving certain roles decision-making power. Assume a Marketing Commander wants to initiate a fund transfer to a newly listed vendor; instead of creating a public proposal, DAO A can set a condition that requires every proposal that belongs to the “Treasury Allocation” category to require approval from 2 Marketing Commanders, 1 Marketing Padawan, and 2 Finance Lieutenants to make it an AP Action. AP Actions mean proposals that are assumed to be approved without the need for everyone voting. These conditions are custom set by the DAO.

The key main differences between public proposals and AP Actions are:

• Public Proposals can either be issued in the general DAO or pods and they can have custom voting distributions like the examples above. They have
• AP Actions have shorter timeframe, can be around 6–12 hours, because it has been approved by the required contributors.

Appointing contributors to a role can be done through AP Actions or Public Proposal, also known as an Election. Using the last example, if the Marketing Commanders have an opening role, they would open an application process for anyone to apply. Then, there are two ways to go about this:

1. AP Action, with a configurable condition set by the DAO, in this case, the Marketing Commanders need to get approval from 2 Marketing Lieutenant members (specified by the DAO) to pass a hiring decision.
2. Public Proposal where the applicants are put against each other in a voting run where all contributors in the Marketing Pod are encouraged to vote.

The clear problem with this moderately hierarchical structure and the AP Actions feature is that the people who elected the so-called magistrates in the first place are held hostage to the elected’s actions and have no power whatsoever to reverse the decision. People are constantly victims of these “fake hopes” that these magistrates promised during the election period, only to underperform or make irrational decisions. The people should have the ultimate power at all times, not the magistrates, since they were elected by the people. Possible solutions to give power back to the people:

1. Using the previous example, DAO A can set conditions for the amount of “against” vote required to veto AP Actions. For example, in the Marketing Pod, AP Actions will get vetoed if there is a majority (51%) against vote in the pod (based on the public proposal voting distribution of the pod) or if it reaches a minimum of 1000 Marketing reputation tokens. These conditions are configurable by the DAO.
2. If elected members take unpopular action or commit any wrongdoing, they will receive negative feedback or have their reputation tokens deducted. This will encourage magistrates to always act in the best interest of the DAO.
3. Allow for re-election or removal from roles if there is a majority vote from the contributors to remove a magistrate. The threshold can be custom set by the DAO, normally for this kind of action it will require somewhere around 80%-90% to remove the possibility of coordinated attacks.

This was inspired by the Roman Empire, where the senate was responsible for setting rules for the Consuls and could veto any decisions made by a dictator. The ultimate power to pass laws lies entirely in the Consuls’ hands, one representing the Patricians and the other for Plebeians; and they can veto each other.

The whole purpose of AP Action is to bring a high level of efficiency while preserving key aspects of decentralization. It’s about creating a non-bureaucratic management and still giving the majority a chance to have their voices heard and veto AP Actions. In addition, for this to work efficiently, every decision taking place inside the DAO or pods needs to be documented. This will create a culture of transparency, which is what DAOs are meant to bring.

Creative twists to AP Actions: Staking or gambling reputation tokens can be a good way to measure the reliability of the AP Action taker. For example, by default, a Marketing Commander should have good intentions behind all their actions, need to consult with the DAO members, and get the required signatures before executing any AP Actions. A DAO can require magistrates to stake 20 Reputation Tokens behind all AP Actions, so if the AP Action passes, the Marketing Commander will receive +20% of the Reputation Token. But if the AP Action gets pushed to the public proposals or falls through, the Marketing Commander will lose the staked Reputation Token because of bad judgment.

Trustware and Socialware Mechanism

In a DAO, there will be certain actions that will be passed through the trustware or the socialware mechanism. Trustware transactions are on-chain and bound by smart contracts, so if a governance threshold is met, the smart contract automatically executes the transaction. Whereas socialware transactions still require manual human action to conduct them. An example of trustware is capital allocation actions, and socialware is like a proposal to remove a member due to misconduct. Trustware and socialware should coexist, and the most successful DAOs will find the balance.

• Metagov Gateway is pushing the concept of trustware by making it easy for anyone to create conditions to execute various actions: https://gateway.metagov.org/

Vote delegation is essential because it removes a responsibility from contributors and allows them to focus on their work. Delegated votes can be removed at any time by the delegator based on the performance of the delegates, and this can be done through analyzing on-chain data. The assigning or election process of delegates that fundamentally enables AP Actions is the prime example of vote delegation. However, vote delegation cannot be a signaling race between contributors, like what happened in MakerDAO. Every contributor should understand the importance and the responsibility of being a vote delegate and act in accordance with the expected behavior.

The Shift to Networked Governance

Pure bureaucracy is too slow and no longer efficient after globalization and the technological revolution took place. This was triggered by the introduction of the neoliberalism concept in the 20th century that emphasizes liberalism and capitalism. This opened up the world and created self-managing capital markets, or what is known as globalization. There is a well-needed shift from the old bureaucratic leadership to a more coordinated governance that is based on collaboration between two or more parties, aka networked governance. The classic example to illustrate the need for networked governance is wicked problems. Suppose there is a high level of crime in a neighborhood that seemed to be a criminal justice problem. However, after further research, most of the crime is conducted by high school students, which shifts the concern to an educational problem. Then, it turns out the kids who are committing these crimes come from single-parent households where the parent is busy working during the day. Now it seems like a problem for the department of social welfare. The takeaway here is that most of the problems in the world are wicked problems, meaning they’re caused by many different factors. What we can do is create a medium or an environment where collaboration between departments is encouraged and fostered; in the example’s case, it would be collaboration between the criminal justice, education, and social welfare departments to solve the high crime problem.

In the next 5–10 years, there will be a Cambrian explosion of DAOs. DAOs will remove all the barriers to collaboration between individuals and coordinated capital allocation, that formerly would’ve taken weeks and months to set up. When this happens, collaboration between DAOs will accelerate the rate of innovation even more and solve the most important interconnected problems humanity faces.

Looking Forward

All the concepts and examples above are meant to show the composability of the proposed system. We should tear down the barriers and build the primitives for DAOs to create governance structures that fit their unique needs. Decentralization is about giving everyone an equal and deserving voice, but this doesn’t mean every DAO should follow the same framework. Governance structure will be a major consideration before contributors decide to participate in a DAO, therefore the process of understanding should be streamlined. That said, it is no surprise if we see a centralized DAO in the future.

Additionally, the composable and forkable nature of blockchain will not only increase the variations but will create highly coordinated communities which will drive innovation like never before. This is because the outcome of DAO’s democracy would be more definite, 70% to 30% or better, unlike the Fosbury Flop phenomenon we are accustomed to seeing in modern democratic structures where the outcome is decided with a 1% margin and half of the population is not in full support. This model is clearly flawed because this thin margin means that every election is decided by a small number of people, which makes it no different than flipping a coin to determine the fate of a country in the next few years. The tighter the margin, the more coercion.

As you may notice, there are a lot of “configurable” words in this article, done to stress the flexibility DAOs will have to configure the best governance structure and management practices that can help them succeed. The challenge here is not to make it complicated and hard to understand, especially for contributors. This modular approach to governance will only create a more vibrant ecosystem, and just like the study of ecology, the higher the biodiversity, the more resilient the system will be. Lastly, we need governance frameworks to reduce the hassle for DAOs from having to start from a blank slate. Ideally, DAOs should be able to see and copy other DAOs’ templates if they wish to do so. This will encourage the sharing of best practices and allow organizations to learn from each other’s successes and failures, further promoting innovation and progress in the DAO ecosystem.

In summary, the future of DAO governance lies in creating flexible, composable structures that can be customized to suit the unique needs of each organization. By adopting a more networked approach to governance, DAOs can better tackle complex, interconnected problems and foster greater collaboration among contributors. As the blockchain ecosystem continues to evolve, so too will the governance frameworks that drive these organizations, ultimately leading to a more vibrant, resilient, and innovative decentralized landscape.