The discussion initiated by Robhaisfield revolved around the risks associated with collateral liquidation in a portfolio containing ETH and USDC. He queried about the asset that gets liquidated when the value of ETH drops. Eboado clarified that the Loan-to-Value (LTV) and liquidation thresholds are a combination of the ones of ETH and USDC, weighted depending on the proportion of each forming the collateral. He further explained that up to 50% of the total value of the collateral gets liquidated, and it's the liquidator who chooses which asset to receive³.

Robhaisfield expressed curiosity about why the system was designed such that the liquidator chooses what gets liquidated, expressing a preference for ranking the collateral himself⁴. Eboado responded that the main reason is the robustness of the system. He explained that liquidation is a critical mechanism of Aave, enabling the 1:1 equivalence between underlying and aToken. He also pointed out that liquidations are influenced by the external market of assets, and having the liquidator choose the collateral to liquidate incentivises them to do liquidations in situations where one of the collaterals is going down and the other not so⁶.

The discussion concluded with Robhaisfield acknowledging the explanation and noting that one of the benefits of using Aave vs. a bank is its programmatically solvent nature⁷. Eboado agreed, adding that the system itself is overcollateralised at any point in time, and the interest rates algorithm creates pressure on liquidity entry when there is not much available for withdrawing/borrowing⁸.