Long and the short of it is I've gotten somewhat involved in the DAO/management side of a certain cryptocurrency. They are collateral in the Rari/Fuse pools, which, as a basic primer are: Pools where you can lend cryptocurrency and others can borrow them. Borrowers get liquidated if their LTV exceeds a certain ratio of their collateral, valuated in USD, and each collateral type has its own respective LTV. So coin ABC might have an LTV of 50%, DEF 65%, XYZ 75%, etc. And a user might borrow X$ of GHI coin. So long as their borrow doesn't exceed the combined LTV of their collateral, they won't get liquidated. A question that's come up in our team is assessing the liquidation risks of certain pools. If there was only one collateral type within a pool, it'd be trivial, since it'd be easy to find the price that one would get liquidated at and compare it against volatility. But since it's a basket, I need a way to compare them as a basket. I'm guessing I could do something like find all the betas to a benchmark cryptocurrency like Ethereum. Then I'd have the beta of all the collaterals combined to use for comparison. This seems simple and straightforward, but just wondering if folks know of better practices.
Start with that. Ideally you want to monitor the covariance of the portfolio, evaluate through time (stressed by events or +/- x%), and determine what levels are appropriate.