The views presented in this article reflect my personal opinions and research, and do not represent the official stance of the Morpho Association or any other organization. They are presented here for public feedback and for informational purposes.

This is the first in a series of articles on decentralized governance, whose goal is to provide background on the problem space and ideas that the Morpho DAO can consider in its own model. I will first consider the challenges faced by governance, synthesizing the literature on this topic. The next article will examine specific implementations, their history, and their success (or lack thereof) in solving the problems described below. Finally, I will apply the ideas of the previous articles in the context of the Morpho Protocol.

There are a few key problems that any system of governance must address:

The principal-agent problem, where the interests of a representative and their constituents are misaligned, is one of the most familiar and fundamental challenges in governance.

The participation incentive problem, where voters have low individual incentive to participate, allowing capture by minorities or preventing quorums from being reached. Low voter turnout and high complexity of governance are problems in both the onchain and offchain worlds.

The succession problem, or the difficulty in ensuring that new generations have the power and skill to effectively govern an organization, including the willingness to continue innovating. Many organizations fail to ensure a robust succession and become “dead players” after the founding generation departs.

The economic capture problem, where there exists an incentive to expend resources to capture the system and extract the value it controls. In onchain systems, the governance system may control assets with greater market value than the circulating supply of tokens, which may be stolen or misappropriated if governance is captured.

In addressing these problems, the reader should keep in mind that “no mechanism exists to guarantee virtuous action”. All great organizations are created by great founders, new generations of strong leaders are required for continued growth, the specific set of individuals and entities in the organization at any given time matters, and no perfect mechanism can replace bad leadership.

This doesn’t mean, of course, that mechanisms of governance are irrelevant. It means that they must be considered taking into account the changing needs of an organization as it evolves. At first, flexibility is paramount and founder control is natural. Later, checks and balances must be introduced, and procedures instituted to help maintain cohesive function in the event of succession.

Let’s address the four problems of governance described above with this perspective in mind.

The principal-agent problem refers to the fundamental conflict of interests between the owners of a company or citizens of a country, and those of the management (in the case of a company) or elected representatives and bureaucrats (in the case of a country). There are often opportunities for representatives to enrich themselves or their allies at the expense of constituents, such as offering preferred negotiating terms to their allies, nepotism in hiring, or even directly stealing from the organization’s assets.

In an onchain context, the principal-agent problem can manifest in familiar ways:

• Founders using their reputation to get access to desirable investment opportunities outside of their core project

• Delegates with conflicts of interest pushing for the selection of certain service providers or technology platforms

• Protocol developers operating a frontend which takes fees from users, separate from the fee switch on the token

• Risk managers making decisions based on social or political influences instead of objective calculations

Ways to address the principal-agent problem:

• Ensure that a robust delegate set exists representing a variety of stakeholder types

• Founders and early investors can delegate their tokens to independent parties, and even enforce that the delegation lasts for a certain duration at a contract level to maximize delegate autonomy

• Ensure that all stakeholders are aligned on a central nexus (ie, no valuable equity separate from a token)

• Harden the protocol by minimizing the scope of changes that can be made

• Add independent oversight of processes (for example, outside financial audits of foundations or labs entities, independent risk ratings)

• Add redundancy, for example, multiple independent organizations contributing to technical development, such as the multiple client teams on Ethereum, or multiple frontend providers in the case of a DeFi platform

A subset of the principal-agent problem is the “accountability problem”, the question of how poorly performing agents can be detected and disciplined or replaced. As discussed below in the “economic capture problem” section, there is no automated mechanism that can reliably detect and respond to the performance of core contributors or delegates. The core solution to this is independent oversight, including a delegate community that has independence from the project’s core leadership, outside financial auditors, and also a healthy media ecosystem that has incentives to surface bad behavior publicly.

Participation in any governance program has costs. Even if gas costs for voting are subsidized or eliminated through a gasless voting scheme, informed governance decisions require a significant attention cost. While Hall and Oak discuss the possibility of incentivizing participation (see footnote 1), representative democracy is the natural equilibrium onchain, just as it is in politics and corporate governance. Except in extreme cases like Optimism, where more than $30m in airdropped OP has been allocated to active DAO voters, DAOs tend to have a small number of voters. Whether it’s Aave, Uniswap, or Lido, it’s rare for any onchain vote to have more than forty direct participants.

Even for a representative system, delegates need some incentive to participate. Either they can be compensated for their efforts and somehow held accountable for outcomes, or they might be entities with a pre-existing incentive to participate who wish to magnify their voice by receiving delegations from others. This topic will be explored at length in my next article, looking at data from several major decentralized finance protocols. My current view is that a healthy delegate set must include a variety of stakeholder types, not exclusively compensated professional delegates. Delegate compensation is sensible, ideally scaled based on a delegate’s voting power.

Decisions that are outside delegates’ scope of expertise can be deferred to service providers or foundation employees, and all governance systems should strive to minimize the number of onchain votes, as well as adopting standards such as a fixed interval between voting periods to reduce overhead for delegates.

Early in an organization’s life, the founders are not mere “agents” representing “principals”, but rather creative directors (and dictators) bringing something new into the world. Too many cooks spoil the broth. In the case of a company, good early stage investors know that the main reason to provide capital into an early stage project is because one trusts in the vision and ability of the founders.

Over time, the founding group is replaced by successors, who may have a smaller direct stake in the organization’s success (smaller ownership percentage). This creates an increased need for checks and accountability to ensure that they do not abuse their power, since there is a bigger gap between their personal interests and those of the stakeholders they serve.

For decentralized and centralized systems alike, among the most critical moments in their lifespan is the handoff of influence from the original founding group. One can observe that many such handoffs in DeFi have been premature, perhaps due to regulatory concerns leading founders to prioritize decentralization over performance. For example, Compound Finance has struggled with voter apathy and failed to compete with Aave and MakerDAO for borrowers in large part because the founders have departed to pursue new ventures. Even in the case of dominant projects like Bitcoin, the lack of a proper successor to Satoshi has prevented the development of useful new features, even those specifically mentioned as desirable by Satoshi, such as onchain privacy.

There’s no simple solution to ensuring that a system remains dynamic and can continue to grow or at least preserve itself after the founding generation departs. The best answer I can give is that founders should strike a balance between establishing credible neutrality and providing a clear vision and direction. Ethereum is a good example of this. Vitalik, the original visionary, remains deeply involved, but the Ethereum foundation has also succeeded in empowering client teams, L2s, and app developers to give input to protocol direction. For example, the Uniswap team was the driving force between EIP-1153, which enabled core features of Uniswap v4. Decentralization does not need to imply that the early leaders have departed, only that new leaders can freely emerge in an open system, with the network’s value driven by the efforts of many independent efforts and use-cases.

It’s common for token governance to have powers such as control over assets worth greater than the protocol market capitalization, or treasury tokens in a quantity greater than the circulating supply. The latter is the reason given by the Optimism Foundation to reject the idea of giving OP holders control over OP treasury distributions. This problem has been a topic of concern among the crypto community for years. In 2021, Vitalik wrote that “all current instantiations of decentralized governance come with great risks… [t]here are two primary types of issues with coin voting that I worry about: (i) inequalities and incentive misalignments even in the absence of attackers, and (ii) outright attacks through various forms of (often obfuscated) vote buying. To the former, there have already been many proposed mitigations (eg. delegation), and there will be more. But the latter is a much more dangerous elephant in the room to which I see no solution within the current coin voting paradigm.”

In my view, there are two major kinds of solutions to this problem:

1. DAO guardian // security council, which consists of giving an elected or appointed council the ability to veto potentially harmful proposals, and/or take other kinds of emergency actions. Such a council might initially consist of core contributors, and later be altered to include elected representatives. Examples of this approach include Compound Finance (the Guardian always had security roles, and was recently extended the ability to veto proposals after a governance attack) and Arbitrum. While not yet working in practice, is Optimism’s bicameral governance, where certain ecosystem actors are given nontransferable voting rights, is also in this category, and should be seen as a broader version of a guardian council.

2. Rate limits and hard caps, for example the Arbitrum DAO can mint new ARB tokens, but at a maximum rate of 2% inflation per year. In theory this limits the ability of insiders to conduct wasteful spending, although in practice the DAO still holds a very large amount of pre-minted ARB tokens and has made some dubious allocations. For another example, the Morpho DAO can activate fees on its lending pools, but there is a hard cap on fees of 25% of the interest paid by borrowers. When compared to the 10-20% average fees on competing lending markets, this ensures that lenders will never face an unreasonable level of extraction.

Vitalik addresses both of these, though I believe gives insufficient attention to the value of a veto-only guardian system in decentralized governance. He brings up a few other classes of solutions, including “Proof of Personhood”, quadratic voting, and skin in the game for voters. I reject the first two, since there is yet no theoretical basis for a non-gameable proof of personhood or quadratic voting system without a centralized identity system.

“Skin in the game for voters” is more interesting, and was one of my central themes in the design of the Credit Guild. This is useful for certain specific classes of decisions, like setting parameters in a lending market or allocating token incentives to a liquidity pool, where outcomes can be easily measured on a definite timespan, but cannot be generalized to couple the outcomes of all governance decisions to the returns of individual voters. One other notable class of systems is vote escrowed tokens, made famous by Curve Finance. Requiring voters to lock up capital does not solve the economic capture problem if the value at risk exceeds the value of the locked tokens. The Curve implementation has also struggled with a lack of skin in the game, where voters are not rewarded more for wise resource allocation, and thus have an incentive to extract the maximum “bribe” revenue that they can. More generally, it is my view that vote-lock systems impose an undue burden on minority voters. Someone who buys into the system cannot vote unless they lock their tokens, and are then stuck with the decisions of the majority, unable to exit.

A robust decentralized governance system should have a strong delegate set. It is highly advisable to have a veto council in order to mitigate economic capture. Governance minimization should be pursued to the extent possible. Founders should balance providing clear vision and direction with a strong commitment to neutrality. This can be done by empowering other voices in the community, including delegates and independent service providers, to provide checks and balances.

Your thoughts are greatly appreciated, both on this article and the upcoming pieces in the series. The next article, coming soon, will examine major governance implementations, weigh their pros and cons, and consider the history of certain (in)famous proposals. Follow on Mirror and on X to stay updated.

Footnotes

1. This problem space is explored in detail in Andrew Hall and Eliza Oak’s paper, What Kinds of Incentives Encourage Participation in Democracy?

2. Samo Burja defines this concept in The Succession Problem

3. Miles Jennings, Machiavelli for DAOs: principles for fixing decentralized governance (part 1)

4. Jeff Amico introduces ideas about delegation neutrality, albeit in the context of VC rather than founder tokens, in Open Sourcing Our Token Delegate Program

5. See, for example, the relay programs on Tally

6. Vitalik Buterin, Moving Beyond Coin Voting Governance

The views presented in this article reflect my personal opinions and research, and do not represent the official stance of the Morpho Association or any other organization. They are presented here for public feedback and for informational purposes.

This is the first in a series of articles on decentralized governance, whose goal is to provide background on the problem space and ideas that the Morpho DAO can consider in its own model. I will first consider the challenges faced by governance, synthesizing the literature on this topic. The next article will examine specific implementations, their history, and their success (or lack thereof) in solving the problems described below. Finally, I will apply the ideas of the previous articles in the context of the Morpho Protocol.

There are a few key problems that any system of governance must address:

The principal-agent problem, where the interests of a representative and their constituents are misaligned, is one of the most familiar and fundamental challenges in governance.

The participation incentive problem, where voters have low individual incentive to participate, allowing capture by minorities or preventing quorums from being reached. Low voter turnout and high complexity of governance are problems in both the onchain and offchain worlds.

The succession problem, or the difficulty in ensuring that new generations have the power and skill to effectively govern an organization, including the willingness to continue innovating. Many organizations fail to ensure a robust succession and become “dead players” after the founding generation departs.

The economic capture problem, where there exists an incentive to expend resources to capture the system and extract the value it controls. In onchain systems, the governance system may control assets with greater market value than the circulating supply of tokens, which may be stolen or misappropriated if governance is captured.

In addressing these problems, the reader should keep in mind that “no mechanism exists to guarantee virtuous action”. All great organizations are created by great founders, new generations of strong leaders are required for continued growth, the specific set of individuals and entities in the organization at any given time matters, and no perfect mechanism can replace bad leadership.

This doesn’t mean, of course, that mechanisms of governance are irrelevant. It means that they must be considered taking into account the changing needs of an organization as it evolves. At first, flexibility is paramount and founder control is natural. Later, checks and balances must be introduced, and procedures instituted to help maintain cohesive function in the event of succession.

Let’s address the four problems of governance described above with this perspective in mind.

The principal-agent problem refers to the fundamental conflict of interests between the owners of a company or citizens of a country, and those of the management (in the case of a company) or elected representatives and bureaucrats (in the case of a country). There are often opportunities for representatives to enrich themselves or their allies at the expense of constituents, such as offering preferred negotiating terms to their allies, nepotism in hiring, or even directly stealing from the organization’s assets.

In an onchain context, the principal-agent problem can manifest in familiar ways:

• Founders using their reputation to get access to desirable investment opportunities outside of their core project

• Delegates with conflicts of interest pushing for the selection of certain service providers or technology platforms

• Protocol developers operating a frontend which takes fees from users, separate from the fee switch on the token

• Risk managers making decisions based on social or political influences instead of objective calculations

Ways to address the principal-agent problem:

• Ensure that a robust delegate set exists representing a variety of stakeholder types

• Founders and early investors can delegate their tokens to independent parties, and even enforce that the delegation lasts for a certain duration at a contract level to maximize delegate autonomy

• Ensure that all stakeholders are aligned on a central nexus (ie, no valuable equity separate from a token)

• Harden the protocol by minimizing the scope of changes that can be made

• Add independent oversight of processes (for example, outside financial audits of foundations or labs entities, independent risk ratings)

• Add redundancy, for example, multiple independent organizations contributing to technical development, such as the multiple client teams on Ethereum, or multiple frontend providers in the case of a DeFi platform

A subset of the principal-agent problem is the “accountability problem”, the question of how poorly performing agents can be detected and disciplined or replaced. As discussed below in the “economic capture problem” section, there is no automated mechanism that can reliably detect and respond to the performance of core contributors or delegates. The core solution to this is independent oversight, including a delegate community that has independence from the project’s core leadership, outside financial auditors, and also a healthy media ecosystem that has incentives to surface bad behavior publicly.

Participation in any governance program has costs. Even if gas costs for voting are subsidized or eliminated through a gasless voting scheme, informed governance decisions require a significant attention cost. While Hall and Oak discuss the possibility of incentivizing participation (see footnote 1), representative democracy is the natural equilibrium onchain, just as it is in politics and corporate governance. Except in extreme cases like Optimism, where more than $30m in airdropped OP has been allocated to active DAO voters, DAOs tend to have a small number of voters. Whether it’s Aave, Uniswap, or Lido, it’s rare for any onchain vote to have more than forty direct participants.

Even for a representative system, delegates need some incentive to participate. Either they can be compensated for their efforts and somehow held accountable for outcomes, or they might be entities with a pre-existing incentive to participate who wish to magnify their voice by receiving delegations from others. This topic will be explored at length in my next article, looking at data from several major decentralized finance protocols. My current view is that a healthy delegate set must include a variety of stakeholder types, not exclusively compensated professional delegates. Delegate compensation is sensible, ideally scaled based on a delegate’s voting power.

Decisions that are outside delegates’ scope of expertise can be deferred to service providers or foundation employees, and all governance systems should strive to minimize the number of onchain votes, as well as adopting standards such as a fixed interval between voting periods to reduce overhead for delegates.

Early in an organization’s life, the founders are not mere “agents” representing “principals”, but rather creative directors (and dictators) bringing something new into the world. Too many cooks spoil the broth. In the case of a company, good early stage investors know that the main reason to provide capital into an early stage project is because one trusts in the vision and ability of the founders.

Over time, the founding group is replaced by successors, who may have a smaller direct stake in the organization’s success (smaller ownership percentage). This creates an increased need for checks and accountability to ensure that they do not abuse their power, since there is a bigger gap between their personal interests and those of the stakeholders they serve.

For decentralized and centralized systems alike, among the most critical moments in their lifespan is the handoff of influence from the original founding group. One can observe that many such handoffs in DeFi have been premature, perhaps due to regulatory concerns leading founders to prioritize decentralization over performance. For example, Compound Finance has struggled with voter apathy and failed to compete with Aave and MakerDAO for borrowers in large part because the founders have departed to pursue new ventures. Even in the case of dominant projects like Bitcoin, the lack of a proper successor to Satoshi has prevented the development of useful new features, even those specifically mentioned as desirable by Satoshi, such as onchain privacy.

There’s no simple solution to ensuring that a system remains dynamic and can continue to grow or at least preserve itself after the founding generation departs. The best answer I can give is that founders should strike a balance between establishing credible neutrality and providing a clear vision and direction. Ethereum is a good example of this. Vitalik, the original visionary, remains deeply involved, but the Ethereum foundation has also succeeded in empowering client teams, L2s, and app developers to give input to protocol direction. For example, the Uniswap team was the driving force between EIP-1153, which enabled core features of Uniswap v4. Decentralization does not need to imply that the early leaders have departed, only that new leaders can freely emerge in an open system, with the network’s value driven by the efforts of many independent efforts and use-cases.

It’s common for token governance to have powers such as control over assets worth greater than the protocol market capitalization, or treasury tokens in a quantity greater than the circulating supply. The latter is the reason given by the Optimism Foundation to reject the idea of giving OP holders control over OP treasury distributions. This problem has been a topic of concern among the crypto community for years. In 2021, Vitalik wrote that “all current instantiations of decentralized governance come with great risks… [t]here are two primary types of issues with coin voting that I worry about: (i) inequalities and incentive misalignments even in the absence of attackers, and (ii) outright attacks through various forms of (often obfuscated) vote buying. To the former, there have already been many proposed mitigations (eg. delegation), and there will be more. But the latter is a much more dangerous elephant in the room to which I see no solution within the current coin voting paradigm.”

In my view, there are two major kinds of solutions to this problem:

1. DAO guardian // security council, which consists of giving an elected or appointed council the ability to veto potentially harmful proposals, and/or take other kinds of emergency actions. Such a council might initially consist of core contributors, and later be altered to include elected representatives. Examples of this approach include Compound Finance (the Guardian always had security roles, and was recently extended the ability to veto proposals after a governance attack) and Arbitrum. While not yet working in practice, is Optimism’s bicameral governance, where certain ecosystem actors are given nontransferable voting rights, is also in this category, and should be seen as a broader version of a guardian council.

2. Rate limits and hard caps, for example the Arbitrum DAO can mint new ARB tokens, but at a maximum rate of 2% inflation per year. In theory this limits the ability of insiders to conduct wasteful spending, although in practice the DAO still holds a very large amount of pre-minted ARB tokens and has made some dubious allocations. For another example, the Morpho DAO can activate fees on its lending pools, but there is a hard cap on fees of 25% of the interest paid by borrowers. When compared to the 10-20% average fees on competing lending markets, this ensures that lenders will never face an unreasonable level of extraction.

Vitalik addresses both of these, though I believe gives insufficient attention to the value of a veto-only guardian system in decentralized governance. He brings up a few other classes of solutions, including “Proof of Personhood”, quadratic voting, and skin in the game for voters. I reject the first two, since there is yet no theoretical basis for a non-gameable proof of personhood or quadratic voting system without a centralized identity system.

“Skin in the game for voters” is more interesting, and was one of my central themes in the design of the Credit Guild. This is useful for certain specific classes of decisions, like setting parameters in a lending market or allocating token incentives to a liquidity pool, where outcomes can be easily measured on a definite timespan, but cannot be generalized to couple the outcomes of all governance decisions to the returns of individual voters. One other notable class of systems is vote escrowed tokens, made famous by Curve Finance. Requiring voters to lock up capital does not solve the economic capture problem if the value at risk exceeds the value of the locked tokens. The Curve implementation has also struggled with a lack of skin in the game, where voters are not rewarded more for wise resource allocation, and thus have an incentive to extract the maximum “bribe” revenue that they can. More generally, it is my view that vote-lock systems impose an undue burden on minority voters. Someone who buys into the system cannot vote unless they lock their tokens, and are then stuck with the decisions of the majority, unable to exit.

A robust decentralized governance system should have a strong delegate set. It is highly advisable to have a veto council in order to mitigate economic capture. Governance minimization should be pursued to the extent possible. Founders should balance providing clear vision and direction with a strong commitment to neutrality. This can be done by empowering other voices in the community, including delegates and independent service providers, to provide checks and balances.

Your thoughts are greatly appreciated, both on this article and the upcoming pieces in the series. The next article, coming soon, will examine major governance implementations, weigh their pros and cons, and consider the history of certain (in)famous proposals. Follow on Mirror and on X to stay updated.

Footnotes

1. This problem space is explored in detail in Andrew Hall and Eliza Oak’s paper, What Kinds of Incentives Encourage Participation in Democracy?

2. Samo Burja defines this concept in The Succession Problem

3. Miles Jennings, Machiavelli for DAOs: principles for fixing decentralized governance (part 1)

4. Jeff Amico introduces ideas about delegation neutrality, albeit in the context of VC rather than founder tokens, in Open Sourcing Our Token Delegate Program

5. See, for example, the relay programs on Tally

6. Vitalik Buterin, Moving Beyond Coin Voting Governance