Summary
We propose listing a USDT Convex leveraged vault on Notional V3 Arbitrum. The vault would allow users to borrow USDT from Notional’s Arbitrum deployment at a fixed or variable rate and LP the proceeds in the Convex boosted Curve USDC.e/USDT pool to earn rewards.
Proposal
We propose listing a USDT leveraged vault using the Convex boosted Curve USDC.e/USDT pool as a strategy. We propose listing the USDT leveraged vault strategy with the following parameters:
| Vault Parameters | Proposed Value |
|---|---|
| LiquidationRate | 1.02 (2%) |
| MinCollateralRatioBPS | 0.13 |
| MaxDevelerageCollateralRatioBPS | 0.19 |
| MaxRequiredAccountCollateralRatioBPS | 1.00 |
| Maximum vault capacity | 300,000 USDT |
| Minimum borrow size | 5,000 USDT |
| feeRate5BPS | 10 (50 BPS) |
| ReserveFeeShare | 80% |
| Strategy Parameters | Proposed Value |
|---|---|
| maxPoolShare | 20% |
| oraclePriceDeviationLimitPercent | 1.00% |
By adding this leveraged vault strategy to Notional, the protocol will effectively accept Convex USDC.e/USDT liquidity pool tokens as collateral. We propose setting the maximum vault capacity at 300,000 USDT to cap the maximum amount of USDT that can be borrowed from Notional to enter that specific strategy. Additionally, we propose setting the maximum pool share at 20% such that users won’t be able to enter the vault if the protocol already holds more than 20% of Curve’s USDC.e/USDT liquidity pool tokens.
We propose setting the maximum leverage at 8.7X such that if a user deposits 10,000 USDT, he can at a maximum borrow 77,000 USDT from Notional to LP in the Curve pool. This maximum leverage ratio (minimum collateral ratio) aims to protect the protocol against a large decrease in the pool’s token value. Setting the strategy’s maximum leverage ratio at 10X ensures the protocol would be protected against a decrease of -9.72% in the pool’s token value relative to the vault’s borrow currency (USDT). Historically, the USDC/USDT worst maximum drawdown over a 1 hour period was -4.3% and -12.2% over a 24H period. It’s important to note that the vault’s maximum leverage ratio is set as a precautionary buffer to protect the protocol and that it is unlikely that the value of pool tokens will decline by such a magnitude relative to users’ debts.
We propose setting the MaxDeleverageCollateralRatio at 0.19, this implies that a liquidator can lower an account’s leverage ratio to 6.3X if it ever breaches the max leverage ratio of 8.7X. This ratio is set such that an account that gets liquidated will subsequently be able to withstand another large decrease in the strategy token price before becoming eligible for liquidation again thereby lowering the risk of an account being liquidated multiple times in a short period of time.
We propose setting the minimum borrow size at 5,000 USDT and the vault liquidation discount at 2%. These parameters aim to maximize the likelihood of successful liquidations under stressed market conditions. To do so, a liquidator’s revenue from the liquidation discount must be sufficient to cover liquidation expenses (gas cost, slippage, Dex fee, price basis). By setting the minimum borrow size at 5,000 USDT and the liquidation discount at 2%, liquidators will likely receive at a minimum 40 USDT. Based on historical market conditions such a liquidation discount will likely cover a liquidator’s expenses under most scenarios.
We propose setting the leveraged vault fee at 0.50% such that borrowers will pay a 0.50% premium on their borrow rate when entering the vault. We propose setting the reserve fee share at 80% such that 80% of the borrow premiums will go to Notional reserves and 20% are paid to nUSDT holders.
We propose setting the oraclePriceDeviationLimitPercent to 1.00%. The oraclePriceDeviationLimitPercent prevents users from entering or exiting the vault if the implied pool prices deviate from the oracle prices by more than 1%. This parameter mitigates oracle attack vectors while being flexible enough to ensure user can enter and exit the vault even if oracle prices slightly deviate from on-chain prices in between updates.