Lending and borrowing protocols form one of the cornerstones of the crypto financial markets. As DeFi has matured, these protocols are typically found at the bottom layer of the “money lego” stack, creating a foundation for building more complex financial instruments. In their simplest form, these protocols serve as a mechanism for retail investors to deposit funds and earn yield on them. Two of the largest lending and borrowing protocols on Ethereum, Compound and Aave, boast a combined ~$21 billion in deposits and have a 35% and 30% utilization rate, respectively. Other protocols tap into these platforms too, with the yield aggregator Yearn Finance, for example, holding ~27% of Compound’s cDai as of this writing.
Since these protocols serve as a foundational layer to the wider ecosystem, it should come as no surprise that these platforms are primarily concerned with risk management. To illustrate this fact, Compound and Aave have partnered with Gauntlet to ensure the risk parameters across their various markets are properly managed. Gauntlet simulates and models a variety of scenarios within DeFi and then proposes updates to parameters (such as an asset’s collateral factor) based on their analysis. The goal is to strike a balance between creating a capital-efficient protocol and mitigating any potential losses for depositors.
Governance plays an important role in these protocols, as risk parameter updates and new asset listings must go through their respective governance process. While this methodology helps ensure the protocol is protected from potential bad debt, it also creates a permissioned process for listing new assets. Long-tail assets effectively have no place in these protocols as they’d be deemed too risky to include in the shared collateral pool.
Rari Capital implemented Fuse pools to cater to these long-tail assets. Fuse pools can be thought of as isolated instances of Compound markets, allowing anyone to spin up and parametrize their own lending and borrowing markets. These Fuse pools have largely been a success, attracting over $900 million in supplied assets, as they’ve allowed larger token holders and DAOs to create markets for their governance tokens.
Although isolated markets like Fuse pools create a buffer between individual lending and borrowing markets and the wider protocol, there is still risk within each market. In January 2022, participants in Rari’s Fuse pool 90 experienced this risk firsthand. The pool uses the FLOAT/USDC Uniswap V3 oracle to retrieve price data for the FLOAT token. An exploiter was able to move the FLOAT price outside of the narrow liquidity band on Uniswap, severely inflating the value of the FLOAT token. With the highly inflated FLOAT value, the exploiter used FLOAT as collateral and was able to drain the majority of the pool’s tokens.
The designs of the aforementioned lending and borrowing platforms demonstrate the trade-offs when creating these markets. Compound and Aave can achieve a higher capital efficiency due to their underlying shared capital pool structure, and to counteract the risks associated with this approach, they must be diligent with their asset listings, relying on governance processes to achieve this. Rari and their Fuse pools isolate risks to individual markets, trading a loss in capital efficiency for a reduction in systemic risks. The FLOAT example, however, serves to reinforce the risks associated with unconstrained collateral usage.
The central problem for a lending and borrowing protocol is to effectively design mechanisms to maximize capital efficiency and provide markets for long-tail assets, all while protecting itself and its users from potential bad debt.
Euler Finance
Euler Finance leverages Uniswap V3’s time-weighted average price (TWAPs) coupled with a unique risk management framework in hopes of tackling the challenges described above. It is currently backed by Paradigm, Lemniscap, and angel investors such as Anthony Sassano, Kain Warwick, and Hasu. The Euler lending and borrowing protocol recently deployed in December 2021 to the Ethereum mainnet and has attracted ~$85 million in deposits, with USDC accounting for ~75% of total deposits.
Permissionless Asset Listings and Risk Management
Lending and borrowing protocols use different approaches to manage assets. Compound and Aave rely on governance to determine which new assets can be added to their protocols, while protocols such as Rari allow the creator of the market to determine the allowed collateral. Euler allows anyone to create a lending market for an asset as long as it has a WETH pair on Uniswap V3, due to the fact that Euler leverages Uniswap’s TWAPs for asset pricing. Utilizing TWAPs for asset pricing implies that Euler markets will be less responsive to and impacted by volatile price movements in the wider market. While this does mean TWAPs are a lagging indicator and can thus be out of sync with certain spot market prices, it also means it will be much more difficult and expensive to manipulate prices.
Allowing anyone to create a lending market around an asset benefits holders of long-tail assets, but it also introduces protocol-wide risks if that asset is exploited, as Euler has implemented a shared capital pool framework similar to Compound. To combat these risks, Euler has designed a tiered risk management framework which assigns each asset to a particular tier.
The majority of assets currently fall into the isolated tier. If a user borrows one of these assets against a certain collateral, then that’s the only asset which they can borrow against that collateral. Currently no assets are listed as cross tier; nonetheless, these assets cannot be used as collateral (this also applies to isolation-tiered assets), but they can be borrowed in tandem with other assets. Collateral-tiered assets garner the most functionality, and they can of course be used as collateral and also borrowed alongside other assets. Assets can be moved from tier to tier via governance proposals.
To further manage the risks within the platform, Euler implements both collateral and borrow factors. Compound, for example, associates a collateral factor with each asset, and this determines the maximum value which can be borrowed against that asset. This collateral factor is independent of the asset being borrowed, ignoring any risks associated with the borrowed asset. With both collateral and borrow factors, Euler is able to account upfront for price action in either direction for borrowed assets.
The FLOAT exploit example not only highlights the risk around collateral allowances, but also shows the risk associated with price oracles. Euler leverages Uniswap V3 TWAPs as an oracle for asset pricing. Currently, for each market setup on Euler, an associated “Oracle Rating” will be calculated and displayed. The framework focuses primarily on the depth and concentration of liquidity for the asset/WETH pair on Uniswap V3.
This methodology will eventually be replaced with a more robust solution. Euler has already developed an oracle tool, which is also open sourced. At its core, the oracle tool simulates how costly it is to manipulate a given token/WETH pair on Uniswap V3. It allows for simulating a variety of scenarios by tweaking parameters such as the TWAP time internal (e.g., 30 minutes) and the target price of the token. Overall, this will provide a much more precise and robust methodology of computing an oracle’s grade.
Liquidation Mechanics
Depending on who is asked, the sentiment around liquidations will likely vary. For traders and borrowers, liquidations can be detrimental to their portfolio and a source of ultimate pain. But liquidations also play an integral role in keeping lending and borrowing markets healthy. If incentives for liquidations are not properly structured, then the worst-case scenario may lead to a protocol accumulating massive amounts of bad debt and potentially never recovering.
In protocols such as Aave and Compound, a position can become subject to liquidation if it drops below a certain health factor. For Compound, this is a function of the defined collateral factor for an asset, and in Aave, users are given a bit of cushion (between 5–20% depending on the asset) once their position passes the acceptable loan-to-value ratio. When a position qualifies for liquidation, up to 50% of the borrowed value can be repaid by a liquidator (this 50% is also known as the “close factor”). By taking on 50% of a user’s liabilities, that liquidator also has a claim to the same amount of collateral value plus a protocol-defined liquidation incentive/discount (a flat 8% in Compound and a range from 4-10% in Aave). This indeed can be a pain point for borrowers, as they can see a substantial amount of their position wiped away in a single transaction. These fixed-rate liquidation incentives and close factors carry the benefit of a more predictable liquidation outcome, but this liquidation design doesn’t allow for the market to have a voice regarding the liquidation dynamics.
Euler has implemented what are referred to as “soft liquidations.” Rather than setting a flat close factor across all markets, liquidators can only remove enough debt and collateral in order to adjust a user’s health factor to 1.25. The health factor is calculated as the ratio of risk-adjusted collateral value to risk-adjusted liabilities. As with Compound and Aave, Euler also has a liquidation incentive/discount. Rather than set these rates at a protocol or market level, Euler leverages a Dutch-auction style mechanism which allows the market to express itself in determining what the appropriate liquidation discount needs to be. The further away a user’s health factor drifts below 1, the larger the liquidation discount becomes (the liquidation discount is capped at 20%). Liquidators therefore can decide when it makes economic sense to carry out a liquidation transaction. Normally, liquidation bots compete with one another in hopes of having their respective liquidation transactions included in the next block, this typically results in contributing to maximal extractable value (MEV). Liquidation bots continue to increase the fees tied to their transactions, resulting in network congestion and increased transaction costs for the average user. Euler’s liquidation design philosophy should help avoid these scenarios where liquidators battle it out in the mempool via priority gas auctions.
Liquidators can also leverage stability pools within Euler which are associated with each market. Instead of outsourcing the upfront capital required for a liquidation event, liquidators can supply funds to a stability pool which can later be used to carry out liquidations. From the liquidators’ perspective, this should generally lower the total transaction costs for liquidations as the capital is sourced within the protocol itself, and there is only an internal price feed to be considered, compared with the scenario of sourcing capital from an external exchange. Additionally, liquidators who provide liquidity to these pools gain the benefit of a “liquidation discount booster,” which effectively increases the profit margins when calculating the total liquidation discount.
Deferred Liquidity Checks
Typically when a user creates a transaction in a lending and borrowing protocol, the system will first perform a health check on the user’s position. If the proposed transaction (for a simple example, this could be borrowing a certain token) would leave their position undercollateralized, then the transaction would not be allowed. This is necessary, of course, to keep the protocol solvent and healthy.
But, imagine being able to bundle a set of actions such that the end result is ultimately a healthy position, but a particular action in the set may not be feasible on its own (e.g., attempting to borrow a token without proper collateralization). The savvy reader may recognize this (potentially complex) transaction description as a flash loan, popularized by Aave. In Euler, the notion of a deferred liquidity check is somewhat of a generalization of the flash loan concept. At a high level, a user is able to build up a transaction and defer the liquidity check on their position until the last action is taken. In protocols such as Aave, a flash loan comes with a small fee (currently set at 0.09%). Euler takes the position that these complex transactions should be free for users, so there is no associated fee when leveraging the deferred liquidity check functionality. Overall, this is a somewhat complex and abstract functionality, but since there are fewer limitations within Euler (i.e., no associated fee structure), users and developers will generally only be restricted by their own creativity.
EUL Tokeneconomics
The protocol’s native token is EUL, which has yet to be released. EUL holders will have the ability to vote on governance proposals (in Euler’s recently announced governance forum), manage the protocol-owned liquidity, reserves, and treasury. There will also be the ability for EUL holders to stake their tokens, which will earn rewards and also serve as a backstop to the protocol.
The max supply of EUL is 27,182,818.284590452353602874, which can be further adjusted after the initial 4 year distribution schedule via governance. The max supply of EUL is of course not an arbitrary value; it’s a symbol and hat tip to Euler’s number which is a widely used constant in many areas of mathematics. The allocation of the tokens will be split between investors, founders and team members, a community treasury, and an initial community distribution mechanism.
Source: Euler XYZ Team
Investor allocations have an 18 month linear vesting schedule with no cliff, and founders have a 48 month linear vesting schedule with no cliff. Team members have a non-linear vesting schedule which follows the pattern of 10% vest after 1 year, 20% of the remaining allocation vests after 2 years, 30% of the remaining allocation vests after 3 years, and the remaining vests after 4 years.
Euler has put together an interesting initial distribution scheme for the community to participate in. The EUL token will be distributed every 100,000 blocks across a variety of markets on the Euler platform. In particular, the distribution will only be on the borrowing side, there will be no distribution for solely supplying assets. The Euler core team will determine the initial set of markets which will be included in this distribution scheme, including USDC, WETH, DAI, and WBTC. The distribution rate will be non-linear over 4 years, with the initial inflation rate set at ~3%, peaking around ~14% after 18 months, then flattening out around ~2.718% at the 4 year mark.
Euler’s distribution scheme will leverage the gauge concept in order to determine the flow of EUL tokens across the various involved markets. EUL holders can stake their EUL against their preferred market in order to drive EUL distribution there, and stakers are not subject to lock-up periods that protocols such as Curve utilize. The gauges will be designed such that EUL tokens are distributed in proportion to the square root of the amount of EUL staked for that market. This function will allow the EUL distribution within a particular market to smooth out once it’s weighted heavier relative to the other markets and will then also disincentivize staking in over-saturated markets.
This approach serves as a more targeted method of EUL token distribution rather than simply distributing EUL to both lenders and borrowers in every market. While lenders won’t be eligible to participate in this EUL distribution, they’ll receive indirect benefits. Borrowing will essentially be subsidized by the EUL distribution, which should increase overall borrowing demand on the platform, leading to increased yield for lenders.
Today and the Road Ahead
Euler currently has ~$85 million total in deposits, ~$34 million in outstanding loans, and has accumulated ~$66 thousand in protocol revenue since its launch in mid-December 2021. If Euler can successfully increase deposits, then this will lower the interest rates for the marginal borrower, ultimately resulting in more protocol revenue. Regarding interest rate models, Euler currently uses the traditional kink model found in both Aave and Compound but has plans to move towards a more algorithmically reactive interest rate framework which will leverage control theory concepts.
As Euler continues to grow, it will also look to diversify both the supply and borrow sides of the protocol. USDC is by far the most supplied and borrowed asset, accounting for ~75% of both the total supplied and borrowed assets. On the USDC supply side, the top 10 suppliers account for ~50% of the total USDC supply, and the top 10 borrowers account for ~66% of borrowed USDC. After USDC, WETH is the most supplied and borrowed asset, making up ~13% on the total supply side and ~18% on the borrow side. The top two WETH suppliers account for ~45% of the total supplied WETH, and the top two borrowers account for ~55% of borrowed WETH.
Being a nascent protocol, Euler will likely look to grow by going after the marginal lender/borrower of longer tail assets, and also by developing integrations with other DeFi protocols. Euler has designed its platform to tokenize supplied assets (i.e., eTokens) as well as debt, in the form of dTokens. Tokenizing debt and allowing it to be transferable is a subtle feature, but it should allow for complex, leveraged positions to be built up which users and protocols may take advantage of. Naturally, Euler will continue competing with protocols such as Rari, Aave, and Compound, and also with new lending and borrowing protocols which are sprouting up across other L1s.
The incumbents will also look to expand their portion of the lending and borrowing market. To this end, Aave V3 was announced towards the end of 2021 and has now been deployed across six networks. While there are numerous new features introduced, the portal and isolation mode additions appear to be most relevant when analyzing the overall design architecture between Aave and Euler. The portal feature will allow Aave to increase its capital efficiency across the various implementations of the protocol. Aave will look to integrate with select bridges, allowing their aTokens to be transferred seamlessly from network to network. In short, this should allow value within Aave to flow more freely between different networks, unifying what is typically considered fragmented liquidity. Aave’s isolation mode effectively combines some of the components of Euler’s asset tier functionality. As new assets are onboarded into Aave, they can be listed in isolation mode. Unlike in Euler, assets in Aave’s isolation mode can be leveraged as collateral, but governance-approved stablecoins are the only borrowable assets and they will also have associated borrow caps.
Although using Uniswap V3 TWAPs for asset pricing allows Euler to develop a logical risk management framework, it also creates a technical dependency on Uniswap V3. Euler currently is only deployed on Ethereum mainnet, which has become increasingly unusable for the average individual due to high transaction costs. Uniswap on the other hand has instances on Ethereum mainnet, Arbitrum, Optimism, and Polygon. Therefore, as Euler grows as a protocol and expands its footprint across other networks, it will already have a critical component of its system in place. Importantly, Euler will not only have to consider where Uniswap V3 is deployed, but there also needs to be significant traction and proper liquidity profiles for assets on these networks.
In order to continue protocol growth, Euler will look to deploy on networks where Uniswap V3 exists and where it is deployed in the future, giving it a first-mover advantage over competing protocols. Given this advantage, Euler may see significant growth, especially because the project launched during a period of rapid decrease in leverage demand. In addition, borrowing will be subsidized when the initial EUL token is distributed. In any case, Euler will continue to develop its product and work towards growing its user base as it seeks to find its place within the lending and borrowing space.
This report was commissioned by Euler, a member of Protocol Services. All content was produced independently by the author(s) and does not necessarily reflect the opinions of Messari, Inc. or the organization that requested the report. Paid membership in the Hub does not influence editorial decision or content. Author(s) may hold cryptocurrencies named in this report and each author is subject to Messari’s Code of Conduct and Insider Trading Policy. Additionally, employees are required to disclose their holdings, which is updated monthly and published here. Crypto projects can commission independent research through Messari Protocol Services. For more details or to join the program, contact hub@messari.io. This report is meant for informational purposes only. It is not meant to serve as investment advice. You should conduct your own research, and consult an independent financial, tax, or legal advisor before making any investment decisions. Past performance of any asset is not indicative of future results. Please see our terms of use for more information.
