On Euler, we use a two-sided approach to estimate risks of borrowing any tokens versus any given collateral. These risks are encapsulated in asset-specific collateral factors (as on Compound) and borrow factors (an Euler innovation).
Consequently, collateral factor reflects the risk of the asset that’s being used as collateral, whilst borrow factors reflect risks of the asset that’s being borrowed.
When it comes to quality of collateral, it is of paramount importance that it ranks very high in our index list.
An illiquid collateral asset can be exploited by causing a price surge to allow a malicious user to borrow an inflated amount of tokens without an incentive to return them.
Alternatively, a collateral asset that’s collapsing in price and is experiencing high slippage makes it uneconomical for liquidators to close positions, leading to bad debts. This scenario is more systemic, which is why only the highest quality assets can become eligible collaterals.
While borrowed assets are less systemic than collateral assets, they also have specific risks.
For example, a borrowed token price that triples in a matter of seconds versus its collateral means there’s no incentive for the borrower to return the token, which results in bad debt. This is why the borrow factor should reflect the volatility and liquidity of the asset.
Alternatively, crashing a borrowed asset's price allows a malicious actor to borrow more tokens than normally possible. This attack can happen if there's not enough liquidity overall, especially if liquidity is too concentrated around a tiny range.