Much of our analysis remains similar to before. At a high level, these recommendations are geared towards the following. We recommend the community to check out our previous analysis for more detail.
Risks associated with increased stablecoin borrowing against LST collateral, amidst lower MATIC LST liquidity
Lower risks of reduced WMATIC supply in the future, which may cause long-term growth risk for Polygon v3
Encourage WMATIC borrowing against MATIC LST collateral
With the Snapshot still underway, we would like to provide our insights and recommendations on this proposal. While we were neutral to the original proposal, at this time, we do not support an additional reduction of MaticX and stMATIC LTs for the following reasons:
The hypothesis that lowering LSTs’ borrowing power will lead to users switching their positions to the underlying assets, as initially suggested, lacks substantial evidence. In fact, it might lead to users withdrawing their funds, resulting in lost revenue for the protocol.
Acknowledging that this analysis is based on a short duration and comes after a relatively minor change, the data gathered in the month following the first phase of LT reduction does not support the initial assumption. The observed trends in token inflows and outflows during this period do not align with the expected outcome.
WMATIC total supply over the past 30 days on V3 Polygon MaticX total supply over the past 30 days on V3 Polygon
stMATIC total supply over the past 30 days on V3 Polygon
Depositors of WMATIC (or any other token) are primarily driven by the potential yield, which depends on the interest rate and borrow demand. Current data from Aave shows a 45.4M WMATIC supply, with only 20% utilized for borrowing. The key to a higher supply lies in increased borrowing demand, which elevates the supply APYs. Therefore, the concern about a low WMATIC supply for E-mode seems overstated - Should the demand for borrowing WMATIC rise, the supply APY will increase, naturally encouraging more deposits.
Aside from the isolated example given in the original post (user 0x1e5b92c66e4cad7963e8dacf1e8d642304c172c8), we haven’t observed a trend of users shifting from WMATIC into LST as collateral. In addition, at this time, we observe the following borrow distribution against the MATIC LSTs:
MaticX
There are currently $42.8M supplied MaticX on Aave V3 Polygon.
A total of ~$16.53M is borrowed against MaticX, predominantly in WMATIC ($13.92M, approximately 84%). Of the WMATIC borrowed, $13.64M is in E-Mode. Roughly $2.89M is borrowed outside of E-Mode.
A total of ~$7.19M is borrowed against stMATIC, mostly in WMATIC ($4M, approximately 55%). About $3.19M is borrowed outside of E-Mode, accounting for roughly 44%.
Finally, our simulations show that the current LT configuration for the LSTs is adequate and can even be further optimized, considering the volatility and liquidity of these assets.
Hey @ChaosLabs, thanks for the feedback here. As a risk manager, Gauntlet’s updates to stMATIC / MaticX LT are meant to be proactive to risk, instead of reactive. The risks we have identified include the following.
It is difficult to see immediate changes in WMATIC supply composition after a short duration and a relatively minor change, as you pointed out. However, just because supply trends have not formed yet, does not mean that there is lower likelihood of them forming in the future.
Thanks for pointing out the depositors of WMATIC are driven by potential yield, which is pretty common for any other token. They are also driven by the ability to borrow against their WMATIC collateral. We also agree that increased borrowing demand elevates supply APY for WMATIC. However, supply APY at the kink for WMATIC, even with the current interest rate regime of slope1 = 6.1%, Uopt at 75%, RF at 20%, is 3.7%, which is lower than the APY for both STMATIC (4.3%) and MaticX (4.7%).
This WMATIC Supply APY will decline to 3% at the kink should AIP-370 pass.
As such, for a user interested in MATIC exposure but looking to borrow stables, it is strictly more profitable to supply stMATIC / MaticX to borrow stablecoins. This may lead to risk in declining WMATIC collateral usage, which is what this proposal is meant to proactively mitigate.
The user that we have collectively noted replaced ~25m WMATIC collateral with stMATIC collateral against stablecoins. This was 35% of the ~70m WMATIC supplied in Polygon v3 at the time. Again, echoing from above, lack of trend today is not indicative of future behavior.
We’ve continuously pointed out the low liquidity levels of Matic LSTs both in the original post as well as here. Based on our Matic LST liquidity data, as of now there are only ~1.4m WMATIC in pools against stMATIC on Balancer, when historically there were > ~$20m TVL in liquidity pools. A similar amount exists for MaticX as well.
A cursory scan of 1inch shows that a 3m sell of MaticX or stMATIC incurs > 30% slippage. Current supply of MaticX and stMATIC on Polygon v3 are each ~40m, implying that ~24m worth in stables could be borrowed in total. As market moves upwards and user risk appetite increases, potentially leading to increased borrowing against MATIC LST collateral, these lower liquidity levels may increase in risk especially when drawdowns occur.
Summary
Gauntlet stands by these parameter recommendations given the analysis above.
