Hello,
This is a proposal for the initial setup for the USDA stablecoin and notably the 3 main modules: Transmuter, borrowing infrastructure and savings solution.
Transmuter
A dual engine
The goal of Angle as a stablecoin in general is to optimize its balance sheet under liquidity and solvency constraints. The stablecoin should be usable as a means of payment, but should also look for optimizing the returns it is generating on its assets in order to build equity for its shareholders.
Here the backing of USDA has been thought as a dual engine to provide through stUSD, its associated savings solution, the best of DeFi and TradFi risk-adjusted yields.
It is conceived as a safe haven where users can park their funds and thanks to the Transmuter technology and its rebalancing mechanisms, they can passively earn the best of DeFi and TradFi yields.
Typically, when yields are higher in DeFi, Transmuter should look to increase its exposure to DeFi yields (within reasonable bounds) and when yields are higher in TradFi, it should look to increase its exposure to real-world assets and tokenized securities.
Handling liquidity constraints
We’ve been working recently with the Steakhouse team on the liquidity maturity of EURA holders. From the conclusions of the work done, what’s essential to achieve is to ensure at any time that there are 30% of liquid assets in the Transmuter to enable anyone to get out.
Even though with Transmuter, users can burn their stablecoins for any of the collateral assets in the backing, not everyone is going to take advantage of this, and regardless of the always open redemptions, we need to take this into account and maintain a liquidity layer for people who may want to exit.
While a higher share of liquid assets like USDC reduces the net interest margin of the protocol, it is also what guarantees that USDA can be minted and burnt with no fees from USDC, or in short that the liquidity of USDA is the same as the liquidity of USDC.
That is why in terms of exposures, in any circumstance, the USDA Transmuter will look to hold a target of 30% in USDC or other similarly liquid assets.
Assets to consider
As a dual engine, Transmuter should look for assets in its backing that enable it to balance between DeFi and TradFi yields.
In the current framework, we’re proposing the following assets:
- USDC: for the liquidity part
- USDM by Mountain protocol: while the stablecoin is not as liquid as USDC, it is still highly liquid and for KYC-ed users, it can be exchanged against USD or USDC with no fees
- bIB01: an ETF by Blackrock investing in short term US government bonds tokenized by Backed (currently yielding 5.24%)
- bERNA: an ETF by Blackrock investing in short term investment grade corporate bonds tokenized by Backed (currently yielding 5.57%)
- steakUSDC: corresponding to deposits in a Metamorpho vault managed by Steakhouse, earning so far the yield paid by USDC borrowers on wstETH-USDC and wBTC-USDC on Morpho Blue markets, extra Morpho and wstETH incentives
We’ll share more soon on the collateral assessment process here (in the meantime you can check Steakhouse website for different collateral overviews), but global policy is to have collaterals with minimal trust assumptions/counterparty/smart contract risks.
On Morpho Blue markets, the idea is to minimize the risk exposure of the protocol and let it perfectly control the assets and collaterals it is exposed to, all while relying on a minimal codebase.
Backed is also one of the main tokenizer introduced here. Backed assets are making up a significant share of the collateral for EURA, they’re setup in a bankruptcy remote way and enable holders to get a claim on the collateral.
Launch setup
While the framework we propose here involves all these tokens, some of the costs in Angle Transmuter are linear in the amount of assets supported.
Typically, when burning USDA, the Transmuter will iterate and read the oracle values associated to all the collateral assets of the protocol to check whether there is a deviation or not.
To minimize these costs, we suggest to start with the 3 following collateral assets: USDC, bIB01 and steakUSDC.
Depending on the evolution of market conditions (e.g did DeFi yields decrease with respect to the yield from RWAs), we could onboard USDM and ERNA in the backing as well.
Transmuter controls its exposures to the various tokens used in its backing, ensuring that:
- it is never at any point overexposed to a token to which it shouldn’t be overexposed
- there are rebalancing mechanisms in place to bring the Transmuter reserves back to their target exposures after some time
For the launch setup, we propose:
- for USDC: No max or min exposure, no mint or no burn fees. This means that anyone can come and mint with an unlimited amount of USDC. The protocol will never block people from minting with USDC (even if it’ll rebalance after large mints)
- for bIB01: max exposure 50%, min exposure at 10%, no mint fees, burn fees at 50 bps to compensate for acquisition costs
- for steakUSDC: max exposure 80%, min exposure 30%, mint fees and burn fees at 5bps to prevent the case of people emptying steakUSDC reserves and permissionlessly unwrapping steakUSDC into USDC in the protocol reserves
The above is what would be enforced onchain and more the security measures in terms of the max that we want to have.
While Transmuter supports negative fees to incentivize rebalances, to avoid any infinite loop or arbitrage opportunities, we propose here to use offchain coordination for the rebalancing and rely on a similar setup to the one that had been voted for EURA.
In terms of target, Transmuter will be targeting:
- 60% exposure to steakUSDC
- 25% to bIB01
- 15% to USDC
Here, steakUSDC is counted as a liquid asset, and so in all conditions there would be >30% exposure to liquid assets with the Transmuter enabling anyone to seamlessly burn for a liquid token with no fees.
When it comes to rebalancing, this is not meant to be binding but here is how a general rebalancing setup could look like:
- if USDC exposure is above 20%, we could rebalance following the desired proportions (25% for bIB01, 60% for steakUSDC) with a priority to increase steakUSDC exposure
- if USDC exposure goes below 10%, we could assume that it’s the steakUSDC layer that is taking the dump and below 50% exposure to steakUSDC we dump some bIB01.
- if bIB01 reaches its max exposure (or are like at 40%), we could rebalance into USDC/steakUSDC, and if bIB01 close to min exposure (at 15%) we could rebalance into it
- if steakUSDC is close to max exposure (~70%), we could rebalance into bIB01 to go back at target, and if steakUSDC is close to min exposure (~50%), we could rebalance into it.
The goal here would be to ensure that the majority of the time, it’d be possible to trade USDC with USDA and conversely with no fees the majority of the time.
Transmuter may freeze burning or charge higher fees when an oracle of one of the assets in the backing deviates from its base value. We define here deviation thresholds below which no deviation is considered.
When it comes to the oracles:
- for USDC: we take a base rate of 1 and apply no deviation when USDC trades plus or minus 5bp with its current price. Typically if USDC trades at 0.9995 or 1.0005, Transmuter considers that USDC is worth $1
- for bIB01: we use Chainlink oracle and then compare this rate with the ATH observed for this token, with no deviation reported when current price is within 20bp of the max price
- for steakUSDC: we combine a Morpho derived oracle and a Chainlink oracle and compare this rate to the ATH observed for steakUSDC, with no deviation reported when current price is within 5bp of the target price
We also propose here to let 100k USDA available on the guardian multisig of the protocol to deal with the rebalancing of the protocol reserves.
These 100k USDA would in fact be an advance by the protocol to the guardian multisig: let’s say there is a rebalancing involving swapping USDC→USDA→steakUSDC, the rebalanced would pay 5bp on the USDA→steakUSDC conversion, but these would go to the protocol. And if it’s the guardian multisig sponsoring the rebalancing transaction, then ultimately, these would be paid back.
Borrowing module
Transmuter is just the tip of the iceberg here and USDA will come as well with its borrowing infrastructure. Contrarily to what is done for EURA, we propose here delegating the borrowing infrastructure on Morpho, in the sense that the protocol will be minting USDA in a Metamorpho vault curated by Gauntlet which will be in charge of managing the risk of the vault.
In terms of collaterals, Gauntlet should be looking for the following assets:
- blue chip ones like wBTC or wstETH
- new LRTs/LSTs which are ready to participate in co-incentives for USDA lenders and for which there is a lot of idle demand to borrow against. Typically, we’ve been looking into accepting Pendle PT tokens as a collateral
On top of the collateral assets (curated by Gauntlet), the key thing to manage here will be the amount of USDA minted by the protocol in such markets.
To avoid situations like the GHO where reserves of the protocol are being emptied and people systematically dump the borrowed USDA thereby affecting the peg, we’ll need to cap the amount minted (equivalent to the debt ceilings we set) by the protocol based on the size of the balance sheet of the protocol and its Transmuter.
If the protocol starts by seeding USDA with 3m USDA through the Transmuter, assuming some liquidity pools, not a full utilization of the USDA lent, protocol can start by lending 3m USDA in it.
In any case, there should be regular proposals to adjust the debt ceilings of these systems.
There are several reasons for why the protocol would be better off delegating its borrowing infrastructure to another system. Typically, this helps for distribution in the sense that everyone on Morpho or the frontends that integrate the system would list Angle vaults (rather than only the Angle Labs frontend), the protocol would be able to onboard other lenders, rather than being the sole lender in this, and it’d be able to work through co-incentives in a single sided way on these vaults thereby reducing its cost of capital.
Savings solution
With this borrowing and Transmuter setup, the protocol will be earning a yield on its reserves, either from the pre-minted USDA on Morpho Blue or on the assets backing USDA from the Transmuter.
Like for EURA, we propose here to launch when USDA goes live a savings solution that is being allocated a portion of the yield earned by the protocol on its reserves. For the computation formula of the yield and APR, we propose to use the same formula as what is currently applied for stEUR.
The savings solution has been deployed on many different chains and to start with, we suggest to activate it on Base, Arbitrum, Optimism and Ethereum.
Note that for the sake of the distribution of USDA, there may be some channel acquisition deals where concluding some fee sharing agreements with large wallets may help for USDA integrations. In due time, we would suggest to amend the stUSD APY computation formula.
An important aspect of the savings solution comes down to how the yield of the assets in the backing is computed. While it’s easy to compute the yield stemming from bIB01 or from the USDA that will be lent on Morpho, the situation is different for the steakUSDC accruing token rewards and not a simple native yield.
Given that the protocol may not always be able to dump the token at their current market price, we propose to estimate the APR to compute the price of wstETH at half of its current market value. This way if the protocol accumulates 1 wstETH of incentives at a price of $3600, it can still afford to dump them at $1800 and remain profitable.
For the MORPHO token, we propose to price them at $0.2 which is inferior to the price during the latest rounds. If the protocol was to get less of this, it would absorb with its equity.
What’s next
Right now, USDA is in beta managed by Angle Labs, we’ll post a proposal and a vote really soon for the DAO to accept the ownership of the smart contracts.
For a breakdown of all the details of how the stablecoin works and its different modules with one another, you can also refer to this blog post.