Frax ETH, A Novel Liquid Staking Derivative Model
Frax’s dual token model achieves higher staking yields than its competitors– How do they do it?
For those still not familiar with the concept, Liquid staking derivatives (LSDs) allow a different way for users to stake their ETH. Unlike the conventional staking contract which requires users to lock up their ETH, LSDs allow users to provide liquidity in a pool and earn both staking and farming rewards. This alternative allows users to withdraw their ETH at any time, without having to wait for the upcoming Ethereum update that will enable staking withdrawals.
One of the newest players in the LSDs market is Frax. This DeFi protocol offers an array of products, including two stablecoins (FRAX + FPI), a decentralized exchange (Fraxswap), a lending protocol (Fraxlend), and a bridge (Fraxferry). With this Frax aims to provide a one-stop shop for DeFi investors, as it covers most of the DeFi primitives available today. The addition of LSDs to the Frax ecosystem further enhances this by providing investors with a way to earn staking rewards while remaining liquid and within the Frax ecosystem. Here is how their frxETH page looks with the rest of the Frax products on the left menu:
Frax’s liquid staking system makes use of two separate tokens: one for staking, and one for farming and withdrawing back to ETH. This allows users to earn rewards for staking while still retaining the ability to transact with the Ethereum network. Additionally, the dual token system allows for more flexibility in terms of staking options, as users can choose to stake a portion of their tokens while still holding some for transactions.
First Token: Staking Yield
sfrxETH (Staked Frax Ether) is eligible for the Ethereum staking yield that comes from validators. It is a interest-bearing token where the exchange rate of frxETH per sfrxETH always increases, so it does not keep a 1:1 peg with frxETH but its price is indexed. For those familiar with the concept, this is the same mechanism as RocketPool’s rETH or Ankr’s aETH. The next diagram shows how this process works:
An ETH holder would deposit ETH in the frxETH Minter and receive an equivalent amount of frxETH. This frxETH can be converted to sfrxETH in a vault and then receive the staking reward, since the exchange rate of sfrxETH to frxETH will constantly increase accruing the staking rewards. Meanwhile the frxETH Minter will send every 32 ETH the deposited ETH in order to be used by Frax’s Ethereum validators.
Second Token: Earning Rewards Without Staking
The entry point to the Frax lend product is the token called frxETH (Frax Ether). The second token is akin to a synthetic and usually trades at parity with ETH. This token is eligible for farming rewards from the Curve + Convex ecosystem, but does not earn staking yield. Frax owns a large amount of CVX, being the second largest holder with 7% of the supply. This allows them to redirect CRV and CVX rewards easily towards the frxETH pool. Extra FXS (governance token of Frax) rewards have been added to the pool through Convex as well. In the next diagram can be seen how would be the process of acquiring frxETH and receiving the rewards would work:
When ETH is deposited in the frxETH Minter, users receive newly minted frxETH. Everytime that anyone deposits ETH in the frxETH minter contract, that ETH is deposited and staked within the Frax Ethereum validators. Due to this reason, the amount of frxETH in circulation always matches with the amount of ETH in Frax’s validators.
Why Two Tokens?
Frax was not the first protocol offering this dual token model: Stakewise and Ankr had a variation of this model before, which seems to have worked well for these protocols looking at their current market share. Nevertheless the leaders of the LSDs market (Lido and RocketPool) were started earlier and still stick to the single token model. The main benefit of this model is that it isolates the volatile token from the earning token. This allows splitting the interest rates between liquidity provision for withdrawal and staking rewards.
This system is more convoluted than the single token model of other LSDs and requires constant monitoring by the team to adjust the farming rewards as needed, but enables a finer degree of control. Controlling the amount of liquidity in the liquid/farming token (frxETH) is critical to allow sfrxETH withdrawals to ETH without impacting on the frxETH-ETH price peg. Increasing rewards would bring an injection of liquidity to the Curve pool in case that it would be required. The ratio of supplies between frxETH and sfrxETH can be seen below:
Historical ratio of total supplies of frxETH and sfrxETH according to IntoTheBlock Research.
Right now about half of the frxETH supply (45.25%) is currently locked as staking ETH in sfrxETH. The rest is mostly deposited in the Curve pool earning Convex rewards. Keeping an eye on this rate can be helpful for risk assessment purposes: a rate close to 100% means that most frxETH is staked as sfrxETH, so there would be less exit liquidity in the Curve pool. While a rate close to 1–10% would be considered safer, since less of the frxETH supply would be used as staking and would be deployed as liquidity in the pool.
Staking Yield Above Market
Currently the staking yield of sfrxETH is at 7.37% APR, higher than the one offered by Lido (4.8%) and RocketPool (4.2%). The reason for this gap is because while most frxETH minted ends up locked in the validator contract, not all the frxETH holders decide to stake. Some of that frxETH is deposited in the Curve’s pool since it yield is more enticing than the sfrxETH staking rewards.
This incentive mechanism allows the sfrxETH holders to have access to the yield that they earn by their stake share plus the one neglected by frxETH holders. This is a ‘positive feedback loop’ that gobbles up liquidity, since the more people move to Curve there’s less sfrxETH so the PoS rewards end up getting divided among less sfrxETH stakers. Therefore the sfrxETH APR goes higher under such a scenario, attracting more capital. The historical returns of both frxETH and sfrxETH and its relationship can be seen in the next image, with the baseline Ethereum staking reward shown in the dashed line:
It can be seen in the chart above how frxETH had usually a higher rate than sfrxETH. Both were trending upwards and stabilizing recently between the 7–9% range. This last month the rates have even out, which makes sense considering that both tokens have similar risk profiles, which is the exit liquidity available in the frxETH-ETH Curve’s pool.
The frxETH ‘Perfect’ Peg
Comparatively to other LSD protocols, so far frxETH has been the one keeping its peg tighter according to the Defiwars page, with a constant and minimal discount compared to its competitors:
This is done because the Frax protocol automatically buys frxETH with ETH when the peg deviates slightly (via their frxETH Curve AMO, Algorithmic Market Operations). This protocol’s liquidity has the advantage of keeping a tight peg and balancing the pool, plus acquiring the farming rewards that will be recycled as bribes to continue bringing incentives to the pool. The source of this ETH buying pressure is a fraction (around 10%) of the initial ETH that is deposited in the frxETH Minter contract. This seems to be a dynamic variable set by the Frax team according to the protocol needs. If they would kept more ETH for these “buybacks” then the staking yield would decrease, and the opposite would happen if fewer ETH would be kept, the frxETH price could not be balanced as much.
With briefs frxETH depegs as low as 0.996, it can be seen that frxETH has kept its price peg very stable. If take a look at a price density chart it can be seen how it tends to stay very flat at 1.000 per ETH, with an average over time on the 0.999 ETH per frxETH:
Summing up, this dual token system allows frxETH to stay one to one pegged to ETH and fully backed by AMOs (protocol owned liquidity) plus staked ETH (sfrxETH). The second version of frxETH is planned for release after Shanghai hardfork, when withdrawals and beacon chain reads are enabled. This would allow external validators to be able to join the Frax validator set. Until that date, the reliance on oracles and multisignature wallets to check validator behavior makes it unfeasible to open the validator set to external validator, according to the Frax team.
