I am in favor in principle here, but I do think that we also need to find a way to quantify “adequate overcollateralization” with the same rigor as the other metrics used throughout this RFC/document. It can’t be a static percentage, as market shocks increase the likelihood of collateral defaults as well as leading to a price decline of RSR. So could we tie it to the basket’s actual risk?
- Estimate what the basket could lose in a serious stress event, including combined failures, e.g. where several positions on the same protocol go down together, not just each one on its own.
- Discount the staked RSR to a stress-adjusted value, i.e. whatever fraction could actually be relied on in a wider market crash, since the price of RSR is highly correlated with overall market sentiment.
This means required OC could be defined as “potential stress loss / stress-adjusted RSR value”, with a fixed floor.
The gist: Changes in the basket may also increase the risk of collateral default and should in turn open a discussion about a new target OC rate. Subsequent proposals might then address this by adapting the revenue shared with stakers by increasing/decreasing their share.
EDIT: I am aware of the complexity of this topic, so a reasonable middle ground would be to at least acknowledge that the “competitive risk-adjusted yield” mentioned in the mandate/methodology should equally apply to staking rewards.