TLDR

Compound

• Compound Comet, or Compound v3 allows Compound to go multichain on EVM-compatible chains. This is important, as Compound is currently confined to Ethereum mainnet which limits its growth and userbase due to high fees.
• Compound v3 simplifies the protocol further. Notably, it introduces the concept of a single borrowable asset, whereby only one asset can be borrowed from each pool. This is to give lenders better control over risk and greater capital efficiency for borrowers - further simplifying the protocol. This is in line with Compound’s focus on providing simple and secure open financial infrastructure.

Maker

• Maker’s reliance on USDC continues to rise, calling into question the value proposition of Dai and its risk profile.
• The protocol is currently in a governance crisis. It has reached a size and complexity that many now believe its current form of decentralized governance has become more of a hindrance than a help.
• There is an internal divide regarding how to resolve the governance issue, with Rune Christensen (Maker founder) proposing a decentralized conglomerate structure of specialized MetaDAOs, and Hasu proposing a simpler, more centralized form of governance.
• Legal ambiguity regarding how a DAO can be governed is a key consideration.
• Despite being a revenue-generating business, Maker has started operating at a loss and likely needs to greatly improve its governance process in order to survive and thrive.

Aave

• Aave recently voted to enter the stablecoin game with GHO.
• Aave may have an advantage over Maker in that it currently has a more centralized decision making and implementation structure that enables it to move faster and innovate. However, with innovation comes additional risk, and Aave V3's cross-chain nature holds bridge risk which has repeatedly proven itself to be very significant.
• Aave V3 was accepted by the community, with points such as a DAO-controlled interest rate pushed back against. Aave has stated that they are taking onboard community feedback before implementation.
• A significant challenge remains for scaling and maintaining the profitability of an overcollateralized stablecoin.

Compound Finance

Compound is an autonomous borrowing and lending protocol built on Ethereum. It can be viewed as infrastructure for developers to build various open financial applications. Compound kicked off DeFi summer in 2020 with the launch of the COMP token but has since fallen from the limelight. However, it is seen as one of the most secure and robust in DeFi and has been widely forked by other protocols across various EVM chains e.g. Bastion on NEAR.

Some key developments with Compound include its institutional-grade treasury product, Compound Treasury. Compound Treasury is a permissioned product available only to accredited institutions. These institutions can participate via on-chain USDC lending or bank transfer to earn 4% APR, with daily liquidations possible.

• Rated by S&P
  • Compound Treasury is the first DeFi product to be given a rating from S&P. It received a rating of B-. Read more here.
• Thai Bank uses Compound Treasury:
  • The venture arm of Siam Commercial Bank (Thailand’s largest bank) deposited an undisclosed amount of funds into Compound Treasury. Read more here.

Compound Comet (Compound v3)

Compound v3 is built to improve both the capital efficiency and risk of the overall protocol. It is defined as “an EVM compatible protocol that enables supplying of crypto assets as collateral in order to borrow the base asset. Accounts can also earn interest by supplying the base asset to the protocol.”

• Compound v3 is proposed to be a version of Compound that can operate on all EVM compatible chains.
• It is designed to be more capital efficient, secure, gas efficient, and easier to govern than previous versions of Compound.
• A key feature is that all pools will have a single borrowable asset, and all collateral factors are tied to this base asset (which will initially be USDC). All other assets in Compound can serve as collateral for borrowing USDC, however, only the single base asset (USDC) is capable of earning yield/being borrowed.

Key Features

• Collateral size limits are introduced for each asset. Governance can limit how much collateral of each type can be used to borrow each base asset. In addition, this gives suppliers of the base asset more control over their risk profile. This could be important as Compound is deployed on other chains and may onboard more risky collateral.
• Different borrowing and liquidation collateral factors are introduced in order to protect borrowers from early liquidation and improve protocol risk management.
  • Borrow Collateral Factor (BCF) is a percentage of the collateral supplied that can be borrowed.
  • The Liquidation Collateral Factor (LCF) is always higher than the BCF. Only when a borrower's balance reaches the LCF does it become liquidatable.
    
    • Risk Dao recently published an analysis that found that a single borrowable asset gives better control over risk for lenders and the opportunity for better capital efficiency for borrowers.
    • Previously, multiple assets could be borrowed against a single collateral type. As each asset is given a collateral factor for liquidations of that asset, then collateral risk parameters must be set for the worst borrowable asset - therefore, hindering capital efficiency.
• Chainlink is used directly as a price oracle as it is portable across EVM chains. This is more efficient than setting up custom price feeds.
• Supply and borrow interest rates can be decoupled, with community governance enforcing all protocol economic policies. Reward systems can be extended by governance in various ways.
• Account management tools added allow other accounts to be given permission to control an account instead of only the private key holder.
  • This can allow a contract's account to be controlled by many users and enables features such as a user having a backup account able to control their hot wallet account etc.

• The liquidation mechanism is separated into 2 different steps

  1. The liquidated debt is repaid from Compound’s reserve while the collateral accrues to Compound’s reserve.
  2. Collateral can be sold until the reserve threshold is reached - thereafter, the buyCollateral function is enabled such that anyone may purchase discounted collateral using the base asset. Once the reserves are above the target reserves, no one may purchase discounted collateral. Liquidations start and end with an absorption of the user's account. If reserves go back below target reserves again, buyCollateral becomes enabled again.

Below is a helpful overview of some of the key differences between Compound v2 and v3.

Source: The State of DeFi Lending by Eitan K

With the introduction of the single borrowable asset, Compound pools will now be standalone, meaning that each market is a separate protocol instance with its own governance configurable liquidation and collateral factors. Today, Compound is as strong as its weakest asset. While Compound has remained very conservative and safe, more degen Compound forks such as Cream have been hacked multiple times due to their more liberal adoption of new assets. With Compound v3, any hacks to these assets will remain isolated to that pool. Compound v3’s new structure enables greater flexibility in money markets across the EVM-compatible chains that they are deployed on.

• Note that this holds the assumption that cross-chain markets will operate efficiently. For example, if a lending market on Ethereum is at 100% utilization, yet the same one on the Compound app on Polygon is only at 50% utilization, Compound will not allocate capital cross chain to balance. This is in the vein of simplification and the difficulties and risks associated with cross-chain activity.

Furthermore, Compound's new account management system does not have cTokens - with the idea to further simplify the protocol and enable greater cross-chain composability.

Business Source License

Compound v3 is under ‘Business Source License’ meaning that protocol governance can deploy new protocol instances and allow other projects the right to use the code. This essentially allows the Compound community to franchise the protocol and leaves the protocol open to use cases not currently envisioned.

Key Takeaways

• Compound v3 is optimizing for simplicity rather than a host of innovative new features. This is in line with Compound’s general focus on being an ultra-reliable and secure protocol. The introduction of a single borrowable asset helps simplify the protocol further.
• Compound’s risk-averse approach is arguably the primary reason for the general lack of attention it receives from the crypto community. However, if institutional adoption of DeFi increases, Compound could be one of the protocols to benefit most due to its safe approach. It exists as a secure borrowing and lending infrastructure and does not go beyond this, despite the growth opportunities elsewhere. Compound’s potential as being a base infrastructure for onboarding institutional capital to DeFi is already being demonstrated, and it is the only protocol to receive a credit rating from S&P.
• The fundamental value of the COMP token is ambiguous, as it lacks any meaningful value accrual mechanism and is used solely for governance. Ideally, Compound can integrate some value accrual properties into the token in the future. However, the lack thereof is in line with the protocol’s conservative nature. Tokens with clear value accrual mechanisms are likely treading on thin ice in relation to securities regulations. It is likely that Compound is awaiting regulatory clarity before incorporating changes that would give the COMP token greater utility/value accrual properties.

Maker

Maker is a borrowing and lending DeFi protocol based on Ethereum. Maker enables overcollateralized borrowing of its native stablecoin Dai by whitelisted assets approved by MakerDAO. MakerDAO is known for its community governance and conservative approach to onboarding new collateral types for Dai. This can be attributed to the MKR token model - in which MKR is minted to compensate LPs if the protocol becomes insolvent. This directly incentivizes MKR holders to remain conservative.

The following graph shows the collateral composition of Dai. Source: Maker Burn as of 8 Aug 2022

As can be seen, the majority of Dai is backed by USDC. Dai has been criticized for its heavy reliance on USDC - and has been referred to as ‘wrapped USDC’ by some of its critics.

Risks of USDC Reliance

In order to maintain peg stability in bear market conditions, Maker has needed to onboard more USDC collateral with the effect of reducing protocol revenues and increasing counterparty risk in Circle.

Examining Circle Risk

• Circle can freeze USDC and can be compelled to do so by regulators. In addition, wallets can be blacklisted, even if they aren’t currently holding USDC. This was recently highlighted by Circle freezing USDC that interacted with Tornado Cash.
  • In the case of Dai, they would need to freeze all or a significant proportion of the USDC in Maker. This means Dai cannot be censored directly, but underlying USDC can be censored. While it would be a drastic measure to freeze all the USDC (and other centralized collateral backing Dai), it is not outside the realm of possibility.
  • A worst-case scenario would see all centralized coins backing Dai freezing their tokens in Maker. This would leave Dai with ~11% ETH and BTC backing (note that this would probably be far less in such drastic market conditions as the price of ETH and BTC would likely fall significantly).
• Due to the existence of some decentralized collateral backing Dai, it arguably somewhat diversifies some risk from directly holding USDC, however, this is minimal and the reality is that Dai is currently heavily reliant on USDC.

Governance

Maker governance has arguably degenerated into a sprawling bureaucracy today, with misaligned interests, gridlock politics, and a lack of accountability.

• On the other hand, one could argue that its arduous governance processes are a feature and not a bug and that the protocol has made good progress in a decentralized manner. It is the third largest protocol on Ethereum by TVL and has a revenue-generating business (although has become unprofitable in recent months). Of course, past and present success by no means guarantees future success and DeFi has become considerably more competitive. As an example, this article also discusses Aave’s proposed stablecoin GHO. Maker’s increased size and complexity are increasingly making the current governance process unsuitable.

There is currently a divide in how Maker approaches solving its governance issues, with both parties recognising the need for a radical overhaul and greater specialisation in decision making. On one side are community members who want to maximize decentralization (with founder and now community member Rune Christiansen as a figurehead) and those in favor of a simpler, more centralized system (with Hasu as a figurehead). Watch a debate between Rune and Hasu on the 9th of August 2022 here.

Rune’s Vision: Endgame

Rune envisions Maker transitioning to a conglomerate design, in which a number of independent MetaDAOs work to contribute to various aspects of the Maker ecosystem. He uses the analogy of Alphabet (Google’s parent company). Maker Governance

• In this, there are various independent companies (MetaDAOs) that all work for the benefit of the parent company (MakerDAO). This would enable specialization across the wider MakerDAO - which would remain simple in itself and focused on growth. Meanwhile, the MetaDAOs would be innovative, fast moving and specialized. They would grow their own unique communities and would be free to innovate and experiment but can be unwound by Maker Core if ineffective.
  • Although Rune uses Alphabet as an analogy - not as a precise explainer of how Maker will work - it is still necessary to compare the two organizations. Arguably, Maker as a conglomerate of MetaDAOs could never achieve the same operational success and coordination as Alphabet - which is ultimately a highly efficient and centralized business. DAOs and MetaDAOs would likely result in coordination challenges and operate at a level of efficiency far lower than that of Alphabet. However, they do have the potential to be far more efficient than Maker in its current state. Also, a stablecoin/synthetic asset protocol does not require the same operational efficiency as Alphabet.
• Rune’s MetaDAO concept would introduce DAO-2-DAO relationships rather than various individual relationships. The rationale for this is to dramatically reduce the number of relationships that need to be managed and keep it to relationships between specialized entities.
  • Each of these MetaDAOs will have their own token - which could exacerbate the coordination issue of a multi-token ecosystem. What are the safeguards to stop a decentralized MetaDAO from turning to Aave or Curve, for example?
  • Beyond coordination difficulties, finding proper utility for these tokens may be difficult, as well as ambiguity around legality.
• Detractors argue that having various MetaDAOs could increase coordination difficulties rather than having units within a single DAO dealing with their specialist area. MetaDAOs may become detached from the general picture of the organization and pursue their own initiative rather than that of a single entity. A solution to DAO governance inefficiencies may not be more DAOs.
• In defence of MetaDAOs, they offer a more decentralized solution to that proposed by Hasu. It may be a more viable option for Maker to maintain MKR’s status as a non-security. However, this remains a point of ambiguity.

Rune also proposes launching a synthetic ETH (MATH) in the short term.

• This is in line with Maker ultimately being a synthetic protocol (Dai is a USD synthetic asset).
• It is proposed that this would be bootstrapped as a Lido Staked ETH wrapper, then upgraded to a fully synthetic asset.
• While MATH fees would be 0% initially, the goal is for it to generate further revenue for the protocol.

Hasu - Simplified (More Centralized) Governance

Hasu is a figurehead for those in the Maker community in favor of a more centralized governance approach. Hasu’s proposal seeks to address the governance issue as well as the general lack of identity and confusion in Maker.

• A key component of this is to create a Maker Constitution that enshrines Maker’s vision and north star, as well as the rights and responsibilities of various stakeholders and Maker’s mandates (e.g. make money for MKR holders, create the most liquid decentralized stablecoin, and become the cheapest source of credit).
  • The constitution would enshrine these principles and strategy and decision-making would be implemented in pursuit of them.
  • A constitution would serve as a guiding principle to which governance decisions can be made and can only be updated by a supermajority vote.
• A ‘Council of Makers’ is proposed - which would appear to operate similar to a board of directors. The council would be charged with setting high level strategy, budget decisions, and managing large DAO transactions e.g. token/debt offerings, buybacks, partnerships etc.

Hasu’s proposed decision making structure would operate as follows: Source: Maker Governance

As demonstrated, this introduces a structure similar to that of a typical corporation for Maker. The constitution enshrines the direction in which all participants must operate and can be changed by a supermajority vote of MKR holders. The specialized council acts in line with the constitution, and council members can be removed at any time by MKR holders. This gives the MKR holders a say in high level governance (changing the constitution, removing Council members etc), but not in its implementation.

Hasu’s vision makes sense from a practical standpoint. However, it is uncertain if this will be compatible with how a DAO can be run from a legal perspective. It appears to operate similar to a corporation (albeit with greater transparency and token holder (shareholder) control). His proposal could represent an improvement of current corporate practices but may be incompatible with what is legally possible with DAOs.

Even if Hasu’s vision was to be implemented, it still carries the strong assumption that a single vision can be agreed upon and defined in a constitution. This would certainly be challenging given the dissonance of the organization at present.

An argument can be made that the reason why there is centralized control in organizations today is because it has proved to be the most effective form of governance. If Maker could be managed by community governance without the need for active involvement then that would be ideal. However, as it has grown inevitably more complex it requires more centralized control. This could be a dilemma most DAOs find themselves in. Maker may not be able to scale or succeed without centralizing. However, if it does centralize governance it could be classified as a security. Regulatory clarity will be required before a definitive approach to governance is taken. If Maker continues growing more complex, more centralized and specialist decision-making endorsed by high-level tokenholder power may be required for it to remain competitive and survive. The path to achieving this remains ambiguous, and it will be worth paying close attention to how this unfolds.

Aave - GHO

Aave has proposed to enter the stablecoin game with an overcollateralized stablecoin called GHO. Dubbed a $1 trillion opportunity by Frax founder Sam Kazemian, it makes sense that a protocol with the size and reputation of Aave would seek to tap into this market. Aave is the third largest protocol in DeFi by TVL after Maker and Lido. The addition of GHO to Aave’s offering will place it in tighter competition with Aave.

• Aave and Maker currently enjoy a mutually beneficial relationship:
  • Maker lends to Aave through the Dai Direct Deposit module providing Maker with a distribution channel for Dai.
  • Aave has access to additional Dai on demand.
• GHO will directly compete with Dai on Aave, and Aave will likely ensure that it will be cheaper to borrow GHO rather than Dai (otherwise nobody would borrow GHO). This will likely result in a decline in Maker revenues (on Aave at least).

GHO

• GHO can be minted by supplying collateral to the Aave protocol. Note that collateral supplied will continue to generate yield for the user, essentially reducing the cost of the GHO loan by the yield earned.
• GHO will initially be secured by assets on Ethereum L1 but will pursue organic adoption on L2s to facilitate wider use of the stablecoin. Aave envisions GHO to be integrated into real-world payments among other things.
  • It will be interesting to see if Aave can achieve this. An uncertain regulatory outlook for stablecoins coupled with the likely reluctance of merchants to adopt a new stablecoin may prove challenging hurdles to overcome.
  • Finding a real, sustainable source of demand will be critical to sustaining GHO.

• All interest payments will accrue to Aave DAO. This is important, as it could generate significant revenue for Aave, which is required for sustainable growth and development and would differentiate Aave from the majority of other DeFi protocols. Maker is widely praised for the cash flow it generates from Dai, and a stablecoin like GHO could make a lot of sense for a protocol like Aave to generate a strong and sustainable revenue stream denominated in a US Dollar equivalent.
  • The counterpoint to this is that revenue will not flow to the Reserve Factor which is arguably important for protocol safety and further growth. Discussion in the Maker community has been around growing the reserve factor to enable the protocol to scale its lending business into other products, as reliance on overcollateralization is limiting and going out the risk curve requires reserves to cover any shortfalls and maintain confidence in the system.
• Interest rates are determined by AaveDAO
  • It will be interesting to see how sustainable this is. Automatically adjusting interest rates may be more desirable to reflect volatile market conditions. Different interest rates may be offered for different risk profile loans e.g. overcollateralized by crypto vs collateralized by real-world assets (RWA).
  • A fixed interest rate is arguably at odds with the Impossible Trinity of currencies whereby a country cannot have all three of:
    1. Pegged exchange rate
    2. Independent monetary policy
    3. Free financial flows.
  • This idea has faced community backlash and Aave has stated that this feedback will be taken into account. Source: Economist

stkAAVE and GHO

stkAAVE holders can avail of a lower borrowing rate on GHO. This is justified as stkAAVE is AAVE staked in the protocol’s safety module which acts as a backstop in the event of a default.

This is an interesting design as:

• stkAAVE holders have their incentives aligned with the protocol and are far less likely to sanction risky activity from the DAO.
  • This is important in light of the various exploits and token failures that have resulted from excessive risk-taking e.g. Cream and LUNA.
• It has the potential to incentivize revenue for the protocol:
  • Lower borrowing rates incentivize the purchase of AAVE to be staked. This should increase buying pressure on AAVE and make the protocol more secure.
  • Lower rates will enable more loans to be taken out and thus more protocol revenue.
  • However, note that DeFi participants that take out loans are still a limited subset of users. Therefore, it is uncertain how profound an effect this will have.

Facilitators

Aave introduces the interesting concept of Facilitators for the issuance of GHO.

Facilitators A protocol or entity whitelisted by Aave governance that is given the ability to trustlessly generate and burn GHO within governance parameters.

The idea is that different facilitators with different competencies will be enabled to generate GHO in different ways. The initial Facilitator will be the Aave Protocol on Ethereum L1.

The proposal envisions multiple Facilitators generating GHO from different sources, e.g. RWAs, treasury backed etc. This shows the innovation made possible by the Facilitator concept as the protocol seeks to find various avenues that can scale GHO beyond what is currently practised with decentralized overcollateralized stablecoins e.g. Dai. Source: Aave Governance

• It is conceivable that protocols such as Balancer will be involved as facilitators of GHO given the close relationship between the DAOs. This DAO-to-DAO potential business model could be lucrative for AAVE given that revenues flow back to Aave Protocol.
• An example of how a potential delta-neutral facilitator could operate would be by minting GHO to provide liquidity to Curve and offset fees with LP yield.
• Members of the Centrifuge core team have expressed their interest in contributing to GHO - and will presumably look to become an RWA facilitator in the future. This would be an extension of their pre-existing partnership with Aave.
  • Centrifuge is well known for working with Maker to finance RWAs on-chain. However, if Centrifuge (and other protocols onboarding RWA on-chain) can find a way to work more efficiently and productively with Aave (due to the current governance issues at Maker), Aave could become the primary DeFi 1.0 protocol for RWA. That said, RWA remain a complex and contentious issue (a discussion beyond the scope of this report).

Cross-Chain Stablecoin Aave has already been deployed across 7 chains. If their portal messaging system is implemented and functions as intended it should enable the easy movement of liquidity across these chains and more. The risks of cross-chain bridging are noted in the section below.

Isolation Mode and GHO Aave v3 introduces Isolation Mode which enables users to borrow stablecoins against isolated collateral. Isolated markets enable Aave to onboard long-tail risky assets as they are isolated from the general Aave money market. This enables Aave to grow its TVL and revenue while isolating risk. Debt ceilings placed on these assets ensure that there is a limit on capital at risk.

Risks and Considerations

• Aave is arguably the most innovative protocol discussed in this report. Out of the 3 protocols, it appears to be the most efficient, with a large team shipping new features at a greater pace than Compound or Maker. This gives Aave an advantage, however, its innovative characteristics also arguably leave it open to more risk.
• Aave will need to remain disciplined in its risk management, which is often at the expense of scalability (an extreme example is UST which scaled extremely quickly in late 2021 up to its collapse in 2022).
• Bridge risk: Aave’s portal feature that allows whitelisted bridges credit lines to make up for cross-chain liquidity shortfalls is a great feature as long as the security assumptions of the bridges hold. Bridges have proved very unsatisfactory when it comes to security over the past year, and blockchains themselves are inherently closed systems.
• In the proposal, it is assumed that during market downturns, GHO demand should increase as the price of collateral falls and users borrow more GHO using stablecoins (in High Efficiency Mode with high Loan-to-Value ratios). However, this assumption is arguably dangerous. While fundamentally different assets, UST rested on the assumption that during market downturns there would be additional demand for UST as well. This obviously did not hold in practice. Liquidity rushed to exit the system and to well-established centralized stablecoins considered safe like USDC.
• Aave has stated its intent on organising hackathons and grants to boost GHO adoption. Like any stablecoin, adoption will be key for it to survive and thrive.
• From a decentralization POV, Aave will ideally avoid a ‘wrapped USDC’ situation whereby the bulk of collateral backing GHO is centralized stablecoins such as USDC. This will be a considerable challenge, and the solution is unclear.
• Smart contract risk is always present, which was demonstrated when Cream’s yUSDvault was exploited for $130m which showed a potential similar vulnerability in Aave as well. Thankfully, Aave quickly removed tokens that posed this risk.
• There is current ambiguity regarding how future regulations may affect stablecoins such as GHO in the future. This should be borne in mind.

Concluding Thoughts

Decentralized Stablecoin Landscape While the future is absolutely uncertain and depends on a large number of variables, not least stablecoin regulation, it is not likely that there will be a large number of stablecoins, with an oligopoly arguably the most likely outcome.

Regulation It appears that a tougher stance on regulation will be taken across the board, with stablecoins especially likely to get regulated given the havoc caused by the UST collapse. It will be interesting to see how this may be applied to stablecoins such as Dai and GHO. Dai remains heavily reliant on USDC which can be frozen at any moment, leading to significant censorship risk.

Furthermore, centralized stablecoins like USDC and USDT have proven much better at scaling and are far more widely used.

Regulatory concerns are also relevant to governance which is discussed below.

Decentralized vs Centralized Governance Compound is an organization that can feasibly remain decentralized as it is focused on being a basic and secure infrastructure. However, more complicated organizations do not have this luxury. Ultimately, Maker and Aave are increasingly complicated and require more centralized decision-making as they integrate real-world assets and launch new products in an increasingly competitive market. As stated above, a more centralized governance structure could also leave a DAO’s token open to being classified as a security.

If GHO is successful and begins to take a significant portion of decentralized stablecoin market share, the impact on Maker will be interesting. It may push Maker to more quickly reform their governance structure in order to remain competitive. It will be interesting to see if it follows Rune's vision or Hasu's more centralized alternative. Its current governance structure, bureaucracy and inefficiencies need to be addressed in order for it to survive in an increasingly competitive environment.