[ARFC] Aave v3 Polygon wMATIC Interest Rate Update


title: [ARFC] Aave v3 Polygon wMATIC Interest Rate Update
Author: @Llamaxyz, @MatthewGraham
Dated: 2022-10-17


Simple Summary

Llama presents a proposal to amend the wMATIC interest rate parameters on the Aave Polygon v3 Liquidity Pool.

Abstract

The current utilization of the wMATIC Reserve is 42.5% with borrowing costs of 6.61%, [1]. As the borrowing cost exceeds the yield obtained from staked MATIC receipt tokens, there is insufficient economic incentive for users to utilize the MATIC eMode.

This proposal is intended to increase the capital efficiency of the wMATIC Reserve by increasing the utilisation of the pool and reducing borrowing costs. This is achieved by amending the Uoptimal parameter and selecting the interest rate at the inflection point on the yield curve to be the lower of the two most prominent staked MATIC receipt tokens.

Utilisation of the liquidity in the Reserve is expected to increase from 45% → 72.5%. At the proposed interest rate inflection point, the recursive stMATIC/wMATIC loop will be slightly unprofitable, with borrowing costs of 6.1% exceeding the 5.9% Proof of Stake (PoS) yield by 20 bps. Borrowing costs are reduced by 12.1% across the inflection points and the DAO’s revenue increases 35.8% at the new breakeven point to current market conditions.

Motivation

With wMATIC borrowing costs exceeding the yield of MaticX and stMATIC, users are unlikely to enter into the Matic eMode without financial incentive. Upon review of the wMATIC reserve, the Uoptimal parameter set at 45% means users can borrow up to 45% of the wMATIC in the Reserve before Slope2 notably increases the borrowing cost.

Increasing the Uoptimal parameter 20% from 45% to 65% would enable users to borrow 20% more of the wMATIC in the Reserve for the same borrowing rate at the inflection point. When utilization exceeds the Slope1, borrowers will be paying the Slope2 interest rate which is set at 300%. Reducing the Slope2 from 300% to 100% in line with other reserves will reduce interest rate volatility.

The breakeven rate on the recursive strategy should occur before the yield inflection point such that rates are less volatile with small fluctuations in utilization. With the Slope1 parameter set at 7%, the recursive yield strategy breaks even at 58.5% utilization. This is 6.5% prior to the inflection point, which is conservative.

The capital efficiency of the Reserve can be improved by reducing the Slope1 parameter to the PoS yield source of the recursive strategy. The PoS yield is currently fluctuating between 5.6% and 5.9%, so changing the Slope1 parameter to 6.1%(-0.9%) changes the breakeven utilization from 54.8% to 62.9% (+8.1%), [2,3].

Based upon the above, there is a 210bps buffer between breakeven on the recursive strategy utilization point relative to the inflection point in the interest rate curve set at 65%. It is therefore possible to increase the Uoptimal parameter and further improve the capital efficiency of the Reserve. With Slope1 set at 6.1% and Uopitmal at 75%, the new breakeven Utilisation is 72.5%(+15.3%) relative to Uoptimal set at 65%. The breakeven point is now 250bps less than the interest rate inflection point.

The graphic below shows the changes in the interest rate.

Specification

The below table shows the current and proposed changes to the wMATIC Reserve.

Parameter Current Proposed
Uoptimal 45% 75%
Base 0% 0%
Slope1 7% 6.1%
Slope2 300% 100%
Reserve Factor 20% 20%
Supply Cap 32.88M 47M
Borrow Cap 9.23M 39.95M

References

[1] Aave - Open Source Liquidity Protocol
[2] Stader | Polygon
[3] https://lido.fi/
[4] Creative Commons — CC0 1.0 Universal

Copyright

Copyright and related rights waived via CC0, [4].

9 Likes

complete support for this proposal, as it’s emerging from the previous discussion on Proposal: Improving wMATIC Parameters on Aave v3 Polygon

if there’s no vocal opposition In this current thread, I think it can be considered mature enough to escalate to snapshot vote stage of the governance process this Friday 21st October.

4 Likes

Life is all about incentive design - especially in crypto.

We are supportive of reestablishing an incentive for borrowing via updating wMATIC reserves. This will contribute more revenue to the DAO which is best to maintain as our ultimate goal.

In favor of expediting this to Snapshot.

3 Likes

Hi Everyone :wave:

A snapshot vote has been created. Voting starts 21/10 and finishes 28/10.

[ARC] Aave v3 Polygon wMATIC Interest Rate Update Snapshot Vote

1 Like

Thanks, @MatthewGraham and Llama team. Gauntlet is conducting analysis on the market risk side and will provide analysis within 2 weeks.

Hi Everyone :wave:

With the recent update from Gauntlet, here, we would like to pick up this proposal and progress towards submitting an AIP.

With the recent publication from Chaos Lab regarding BorrowCaps, we are proposing adjusting the BorrowCap risk parameter in response to increasing the Uoptimal from 45% to 75%.

Parameter Current Proposed
Uoptimal 45% 75%
Base 0% 0%
Slope1 7% 6.1%
Slope2 300% 100%
Reserve Factor 20% 20%
Supply Cap 32.88M 49.32M
Borrow Cap 9.23M 27.95M

Level 1 Borrow Cap = SupplyCap * (UOptimal+0.1) = 32.88M * (0.75+0.1) = 27.95M
Level 2 Borrow Cap = 0.7 * CurrentSupply = 0.7 * 17.54M = $12.28M
The recommended borrow cap should be the larger value between Level 1 and Level 2.

With Liquidity Mining likely to occur on this Reserve and the known directional affects this will have on Supply / Demand, we prefer to also utilise the upper band outlined above.

In setting the Supply Cap, current utilisation 53.18%, we expect more than $13.94M of inbound deposits once Liquidity Mining commences. Therefore, we would like to suggest increasing the Supply Cap by around 50%, to 49.32M. This would be enough for all of the wMATIC, 26.60M, on v2 to migrate plus an allowance for some extra liquidity. Given the sizeable increase in BorrowCaps already suggested, we suggest waiting to see how the market reacts to Liquidity Mining emissions before making further adjustments.

If any risk parameters suggestions from the community were to emerge, we will gladly update our proposal. Given the introduction of SupplyCap and BorrowCap changes, another Snapshot vote to confirm the path forward is to be performed.

If the community wants to explore enabling Stable Borrowing of wMATIC, then this would be an ideal opportunity to implement such a change. Let us know in the comments if there is any appetite for such a change and we will update the post accordingly.

1 Like

I highly recommend to also determine caps in token amount not only $amount.

On the sc the configuration is in token amount, so when a snapshot vote agrees on x$ it will be impossible for you to implement the snapshot in your proposal making this unnecessary inaccurate.

I’m in favour of the supply cap increase.
I don’t understand the level1/level2 borrow cap - what does this mean? I think the formula provided in level 1 makes kind of sense.

We amend the cap values to be quantity of tokens rather than $ amount.

For more information on the level1/level2 borrow cap, see this link.

More generally, we have been in touch with Chaos and we are hoping to have feedback in the coming week. Hopefully we ship this proposal before the christmas holidays. :slight_smile:

1 Like

The proposed interest rate at the optimal point is lower than the current APR of stMATIC (6.3%). As such, Gauntlet has the following concerns regarding wMATIC lending pools:

  • Lending Pool utilization above the optimal point will lead to more volatile interest rates. The user experience on Aave could be negatively impacted if interest rates are volatile to small changes in pool utilizations.
  • High utilization might prevent suppliers from withdrawing collateral and increase the risk of failed atomic liquidations.

In consideration of the above factors, Gauntlet recommends changing the slope 1 rate to 6.5% to set utilization below the optimal point.

2 Likes

Hi @Pauljlei,

Originally when we published this proposal, we reached out to Lido and were advised the yield on stMATIC post fees was fluctuating between 5.6% and 5.9%. We have since reached out, and can advise there is a discrepancy between Lido, Stader and Polygon Foundations front end. The earlier indicates 6.3% in line with your comment, Stader is reporting 5.76% and the Polygon Foundation indicates 5.0843%. The Lido front end is hard coded and set at 6.3%.

Based upon the latest update, we are comfortable using 6.1%, 20 bps higher than the range advised in the original post as the preferred Slope 1 parameter. 6.1% exceeds the 5.76% offer by Stader and is considerably higher than the 5.0843% advised by the Polygon Foundation.

1 Like

Chaos Labs has been working with Llama on supply and borrow cap amendments for this proposal.

After analyzing wMATIC, we recommend amending the supply and borrow caps as follows:

  • Supply Cap - 47M MATIC
  • Borrow Cap - 39.95M MATIC (utilizing the borrow cap methodology with the updated supply cap and Uoptimal point)

The analysis has been done using our new supply cap methodology, which we aim to share with the community early next week.

2 Likes

Thank you very much @ChaosLabs for help on the SupplyCap and BorrowCap recommendations:

We have updated our proposal to reflect the below:

Parameter Current Proposed
Uoptimal 45% 75%
Base 0% 0%
Slope1 7% 6.1%
Slope2 300% 100%
Reserve Factor 20% 20%
Supply Cap 32.88M 47.00M
Borrow Cap 9.23M 39.95M

The BorrowCap now uses the new SupplyCap as an input where as the previous post used the current SupplyCap. This increased the BorrowCap from 27.95M to 39.95M.

We will proceed to create a Snapshot to commence Monday with an AIP to follow early in the new year.

https://snapshot.org/#/aave.eth/proposal/0xb1086fea9d6f4c2fcde0362768e4cd8c2bc1a42a1897704ac5639490f1ef895e

Hi Everyone :wave:

A quick update on this proposal. Llama has asked @ChaosLabs to reconfirm the SupplyCap and BorrowCap for wMATIC and late yesterday, we also submitted the payload to @bgdlabs for peer review. The normal review cycle is 2-3 days.

Ideally this proposal is submitted for voting later this week and if not, then early next week.

Once this upgrade has been implemented, it is likely that rewards from various partners will start soon there after.

1 Like