This is a crucial discussion about LSTs, and within the ACI, we view the Polygon PoS as a “canary” network for LSTs in general.
Both MaticX and stMatic can be redeemed at fair primary market value on ETH L1 within three days.
Aave V3 market utilizes calculated price feeds that track the primary market pricing for both assets.
In terms of liquidity, both assets are essentially MATIC with a three-day delay, and their secondary liquidity is not as significant as other assets. Their circulating supply is elastic, with anyone able to mint and burn new assets on L1 within a reasonable timeframe.
That’s why, for both LST assets, liquidity risk and especially secondary liquidity risk are less relevant.
The actual risks with these two assets involve protocol failure risks from Lido and Stader, as well as the L1 <> L2 canonical bridge failure. These risks should be the primary focus of the risk team when modeling risk parameters for LST tokens on Polygon PoS. Other risk parameters should be considered within the scope of MATIC itself since LSTs are MATIC wrappers with minimal timeframe delays.
There is simply no scenario in which the Aave protocol incurs bad debt if, for some reason, MaticX’s secondary price on PoS becomes depegged, because the protocol takes primary market prices into account. If such a scenario occurs, arbitrageurs would profit by leveraging on Aave and other markets on PoS, and/or by purchasing LSTs on PoS and redeeming them for fair value MATIC on L1 with a slight delay.
We believe it is essential to have this conversation now since, after the Shanghai upgrade (in 8 days) and the withdrawal queue clearing (a few weeks/months after Shanghai), ETH and its LSTs will exhibit similar properties.
LSTs represent a significant source of growth and revenue for the Aave protocol, and risk teams should not hinder Aave’s growth by employing inadequate models. Protecting the protocol is everyone’s priority, but let’s do so with the right data by implementing the appropriate models.
We invite both risk service providers to update their models concerning LSTs in general.