Proposal: Improving wMATIC Parameters on Aave v3 Polygon

[YJN58]

Summary:
Optimizing wMATIC interest rate model to meet the borrowing demand of wMATIC generated from the inclusion of stMATIC and MATICx as collaterals and increasing their respective supply caps on Aave v3 (Polygon mainnet).

Motivation:
Similar to how the demand for ETH borrowing increased ~ 4-6 times previous levels after the inclusion of stETH on Aave v2 (Ethereum mainnet), the borrowing demand of wMATIC will potentially increase with the inclusion of stMATIC and MATICx on Aave v3 (Polygon mainnet) and more so with accompanying E-Modes. Currently, stMATIC and MATICx earn ~5.7% APR and ~5.3% APR respectively from staking rewards, while still being used as collaterals on Aave v3, allowing users to go up to 20x leverage theoretically on both stMATIC and MATICx while borrowing wMATIC under new E-Modes (max LTV: 92.50%, liquidation threshold: 95%). So, it is now critical to optimize the current interest model of wMATIC as it will allow Aave v3 to get potentially more revenue, more new users, wMATIC lenders will receive much better returns and possibly Aave v3 can become the primary liquidity sink for MATIC on Polygon PoS. Further integrations such as leverage farming via instadapp will serve as an additional catalyst towards attracting more liquidity and adoption.

Rationale:

Problems with current wMATIC rates on Aave v3:
The current staking rewards of stMATIC and MATICx are ~5.7% APR and ~5.3% APR respectively while the borrowing rate of wMATIC is 6.23% APR currently because of the low OPTIMAL_UTILIZATION_RATE of 45% and the utilization rate is already close to the ”kink”. There is incentive for recursive borrowing only if the staking rewards % of stMATIC and MATICx are greater than the borrowing rate of wMATIC. So, it is suggested that the borrowing rate should be around 4.80% APR for borrowers to have good benefits. This would require the IR model to be optimized in a way where borrowers have enough incentives to borrow and lenders have better incentives for lending.

So, the following updates are proposed to the wMATIC IR model as the current model is not efficient enough as suggested by the impact analysis of the revised model and risk assessment of wMATIC.

• OPTIMAL_UTILIZATION_RATE = 70 (previously: 45)
• baseVariableBorrowRate = 0 (previously: 0)
• variableRateSlope1 = 4.80 (previously: 7)
• variableRateSlope2 = 100 (previously: 300)
• stableRateSlope1 = 7 (previously: 9)
• stableRateSlope2 = 100 (previously: 100)
• reserveFactor = 10 (previously: 20)

Impact of revised IR model:

Projected rates with the current IR model at 4.80 % borrowing rate:

• Utilization: 30.85% [low utilization leading to less revenue generation]
• Borrowing rate: 4.80%
• Supplying rate: 1.18% [low incentives for new lenders to come in]

Projected rates with the proposed IR model at 4.80 % borrowing rate:

• Utilization: 70% [more revenue for Aave]
• Borrowing rate: 4.80%
• Supplying rate: 3.02% [much better rates for new lenders to come in or to get a discount on stable coins borrowing against MATIC]
• Variable borrowing rates (% APR) of stable coins on Aave v3 (Polygon mainnet) as of 24/09/2022.

USDC	1.64%
USDT	3.10%
DAI	2.05%
MAI	2.76%
agEUR	2.02%
jEUR	3.73%

Updated wMATIC IR Curve

Risk Assessment:
MATIC token has now been listed over 60 major exchanges globally, and is trading against more than 50 base pairs. In terms of market cap MATIC is valued at $5.50B and FDV of $7.40B. 24H trading volumes of >$315M make it one of top 20 traded coins across DEXs and CEXs (source). The total on chain liquidity of wMATIC on Polygon PoS is $122M currently with 1M daily average volume of $1.34B which is on par with the numbers of wETH on Polygon (source).

The matrix below shows the figures used to quantify MATIC’s risks per factor.

(source1, source2)

The matrix below shows the risk ratings per factor based on Aave’s risk assessment methodology.

(source1, source2)

We can clearly see that MATIC’s market risk parameters have significantly improved over the last one year which strongly supports the proposal to adopt the new interest rate model.

Also, a recent example of improvement in the risk parameters of wMATIC can be seen with QIP084 of QiDao Protocol which is Polygon’s native stablecoin issuance market where they decided to increase the minimum CDR to 120% from 150% i.e.~83% LTV as they believed that the previous maximum LTV of 66% did not reflect the improved risk profile of wMATIC.

Supply Cap of stMATIC and MATICx:
The total MATIC staked on Lido (stMATIC) and Stader (MATICx) currently is ~47.90M and ~40.20M respectively and to support the expected increase in their deposits on Aave v3 as a result of improved yields for the users due to recursive borrowing under E-Mode with improved IR curve, the following updates are proposed to increase the supply cap of the two liquid staked MATIC tokens.

stMATIC = 15M (previously 7.5M)
MATICx = 15M (previously 6.0M)

Conclusion:
Refactor the wMATIC interest rate model as the borrowing demand will potentially increase and market risk parameters of MATIC have improved. The updated IR curve will make it better for the borrowers, lenders and increase the protocol revenue from wMATIC by ~2.5x (assuming increase in wMATIC supply to fully utilize the stMATIC and MATICx collaterals under E-Modes and borrowing rate at 4.80% APR)

Specifications
Author is a member of the Polygon team.

I appreciate you taking the time to review. Please let me know if you have any questions or concerns.

• Improve wMATIC paramters
• Further discussion required

[Benjamin918]

Great proposal ser!

MATIC has proven to be a reliable asset for lending. We’ve had no issues at QiDao with MATIC since updating the max LTV, even through the most recent market crash in May. As detailed in our community risk matrix, onchain liquidity and slippage figures for MATIC have also been quite healthy.

I think that this move will benefit Aave, as it would bring more usage & fees within the scope of what is safe.

[CBeckman]

There are a lot of pivotal key points in this proposal. To follow @Benjamin918 's reply, Matic indeed proved its reliability within the last 12 months, and now deserves to be unleashed.

As stETH’s ecosystem was incremental to Ethereum’s intrinsic growth, Matic’s staking derivatives hold the same potential given a proper expansion model. However, the current parameters, if left untouched, will rapidly incur some forced limitations which will ultimately restrain potential short-long terms newcomers inflow. Not that AAVE is entirely responsible, but AAVE is King, and this proposal is one serious order that has to be considered solemnly!

This proposed model seems to me like a solid short-term solution that might resolve a long-term problem.

Global Matic addresses
…

ETH Matic addresses

Currently, over 90% of Matic live outside of Polygon’s ecosystem (where 33% of this 90% is in POS staking contract), and this number keeps on growing as time goes on. “What’s the point of building a grand Castle if you and your family have to live in the lodge next door?”

Joke aside, there are many reasons why this is so. One of them is the lack of liquidity for large bag holders. This vicious circle is certainly not easy to rupture, but every step in the right direction may, at any time, create a momentum effect and bring some of these precious Matic back.

I strongly suspect that by implementing the proposed changes, the liquidity of AAVE/Polygon/Dex will deepen alongside the growth of liquid staking scenarios (e.g. Leveraged Staking) similar to what happened on Ethereum after stETH’s release, especially considering that the proposed model follows ETH’s parameter on AAVE V2.

[Racoon]

Definitely agree with @Benjamin918 and @CBeckman

Improving the Parameters of wMATIC will have a positive feedback loop effect on the Aave v3 Polygon market and will also help migrate liquidity from Aave v2 to Aave v3 due to market inefficiency of v2 as a result of improved parameters of wMATIC on Aave v3.

The flywheel effect:

1. wMATIC lenders from Aave v2 would migrate liquidity to v3 due to improved parameters and increased demand for wMATIC leverage which would increase the utilization ratio and hence the supply APR.
2. As a result of Increased supply APR (3.02%), wMATIC lender can now borrow stablecoins at a discount.
3. As the demand for stablecoin leverage increases in the v3 market, lenders from the v2 market will be incentivized to migrate liquidity to earn higher supply APRs lending through lending stables.
4. More stables in Aave v3 = Higher leverage using Aave’s e-mode = higher protocol revenue and supply APR for stablecoin lenders
5. Improved parameters of wMATIC would help kickstart e-mode for stMATIC and wMATIC and will deepen liquidity and also greater adoption of liquid staked derivatives of MATIC which would help in further decentralization and capital efficiency.

The net result of this is extremely positive for both Aave and the entire Polygon ecosystem so I am upvoting this brilliant proposal

[Kyros]

Amazing and Well detailed proposal . Matic has proven itself to be a valuable asset to the aave ecosystem . This move will benefit aave in many ways to come

[Vibin]

Thanks for this proposal. At Stader, we believe this would be a critical step towards bringing down Matic borrow rates and making popular strategies like Leverage Staking possible. More importantly, this could unlock the power of liquid staking and make more of the ~3.1 bn Matic staked accessible in the Defi ecosystem.
@Racoon and @CBeckman have both articulated the potential of this move and we wanted to add some colour to that line of thought.
Based on observations on Ethereum and other ecosystems, it is apparent that leveraged staking and higher utility of liquid staked tokens as collateral can act as a catalyst for an increase in (1) liquid staked assets and (2) their deployment in lending protocols like Aave and broader DeFi. Currently,

1. More than 30% of staked eth is liquid staked vs <3% of staked Matic being liquid staked. Thus making a large % of MATIC inaccessible to DeFi on Polygon
2. Nearly 25% of liquid staked derivatives on ETH is deployed as collateral on Aave ETH market of which nearly half the collateral is used for leverage staking and the rest for borrowing other tokens in a more capital efficient manner. This number is insignificant currently for liquid staked Matic tokens

We support this proposal and look forward to seeing what this change can do for the Polygon, AAVE and Stader communities.

[Simsala]

Thank you very much for this proposal!

This should be a catalyst in adjusting the borrow rates for Matic, which helps not only Matic but Liquid Staking solutions, and opens them for more strategies all over DeFi.

Hard yes.

[MatthewGraham]

Hi @YJN58,

Thank you for bringing this proposal to the forum. I think there have been many conversations around the potential recursive strategies on Aave v3 on Polygon. These strategies were flagged during the listing of stMATIC and MaticX.

Reading over the proposal, it seems we have a mix of the following:

• Yield Curve parameters for wMATIC
• Enable Stable Rates for wMATIC
• Increase LTV and Liquidation threshold for wMATIC ?? (not sure if this is proposed or not)
• SupplyCap increase for stMATIC
• SupplyCap increase for MaticX

This is a lot of changes for a single proposal and voters can only endorse all or none of the changes. I believe it is best practice to narrow the scope of the proposal to remove the conflict of passing all or no upgrades.

That said, the merit is great and directionally I like where this is heading.

Having recently performed the risk analysis on both MaticX and stMATIC, I think changing both Reserves to a SupplyCap of $15M is ambitious. Currently, there is not $600K in deposits between them. Although this is expected to increase, there is a track record of increasing SupplyCaps when the market suggests it is needed. At the moment the market is not signalling it is needed.

Both stMatic and MaticX use Calculated Oracles which do not reflect the spot pricing which Liquidators are expected to use when liquidating bad debt. Aave v3 is a safer design than v2, but does not yet have Safety Module coverage and any deviation between the calculated oracle and spot pricing represents a risk to Aave. Can you please outline how with such concentrated liquidate, partially dependent on liquidity mining and several DEX pools migrating liquidity how now is the right time to introduce more risk given the calculated oracle exposure ?

I noticed the proposal utilises the older style thinking toward listing assets and I have used it in the past when I with Index Coop’s support changed the wMATIC parameters to enable MATIC2x-FLI and iMATIC-FLI products. This is not applicable to the rate curve changes, so I am assuming there is a LTV and Liquidation parameter change in this post.

I believe rather than framing the risk analysis how it has been in this proposal, analysis more tailored to amending risk parameters is needed. Volatility is lower which supports this effort. However, I think we need to look at the following parameters:

• Value at Risk based on say a 50% price drop
• Collateralisation Ratios of the top 30 Holders
• How the wMATIC is being used

Example: Aave - Open Source Liquidity Protocol

I would note the wMATIC LTV and Liquidation Threshold parameters do not affect the recursive yield extraction for depositing stMATIC and borrowing wMATIC etc… It would be great if these parameters are being changed that they be clearly stated. Perhaps I missed this though.

I support the wMATIC yield curve changes and introducing the Stable Borrowing functionality. But I don’t support amending Risk Parameters post listing without deeper risk analysis. I would like to see three separate proposals be created:

Proposal #1 - wMATIC

• Enable Stable Borrowing
• Increase LTV
• Increase Liquidation Threshold
• Reduce Yield Curve (Borrowing Rates)

Proposal #2 - stMATIC

• Increase SupplyCap stMATIC

Proposal #3- MaticX

• Increase SupplyCap MaticX

I am happy to cast a YAE vote on some aspects of this proposal whilst we improve on the risk analysis. I would like to see some commitment from someone to monitor and maintain these parameters as Gauntlet is not yet covering Aave v3 on Polygon. I think this is an important feature to consider adding to the proposal.

May I suggest waiting to include the new oracle from Chainlink that uses spot based pricing for stMATIC on proposal 2. I’m not sure how far away the MaticX spot pricing based oracle is. It is safer to keep these markets smaller whilst we use a calculated oracle.

I would also consider increasing the wMATIC yield curve on Aave v2 within Proposal #1 or in a new proposal. Increasing borrowing costs on the v2 liquidity pool further encourages Users to migrate to v3. Let me know if that is something we can do separately to this proposal.

[Marin]

This is a step in the right direction for sure.

[Racoon]

There is close to ~92M MATIC staked cumulatively with LIDO and Stader. The low number of $600k is because the current yield curve does not make sense for people to use e-mode, so they would rather use Balancer or Curve instead to provide liquidity.

If the parameters of the yield curve and the supply cap is increased, the flood gates will open

[YJN58]

Hi @MatthewGraham

We are extremely appreciative for the detailed feedback on our proposal and would like to clarify a few things to the Aave and Polygon community:

The intention of this proposal was to build an initiative to unlock the power of liquid staked $MATIC via LIDO and Stader Labs by working together with Aave. These suggestions go hand in hand and would result in a win-win for both Aave and Polygon’s DeFi ecosystem as agreed by several members of the forum. I would like to thank the Aave community for their positive feedback for our proposal.

Although the proposal seems to ask for a lot, the scope is much narrower since we have not requested to alter the risk parameters like maximum LTV and liquidation threshold. The two major asks are:

1. Optimisation of the interest rate curve for wMATIC
2. Increasing the supply cap to 15M for both MATICx and stMATIC.

Regarding the migration of liquidity from Aave v3, by simply changing the parameters of wMATIC, it would create a positive feedback loop as mentioned by @Racoon that would see liquidity flow into the Aave v3 markets:

Below is an special Aave french chart which explains the entire positive feedback loop in a fun and easy way:

Regarding the current size of the stMATIC and MATICx markets on Aave v3 being small, @Racoon has pointed out very clearly as to why such is the case. With the current rate curve the borrow APR is ~6% which is greater than the staking APR for both stMATIC and MATICx. Since one can only borrow wMATIC against both assets using Aave’s e-mode, the size of the market remains at $600k. With the implementation of the new rate curve, products like Instadapp’s leverage strategy will definitely pave the way for a lot of new capital to flow in and will increase the utilization of e-mode which means more protocol revenue for Aave.

We are also confident that the Stader Labs and LIDO team should be able to address all the concerns raised about the calculated price feed and also liquidation risks in an extremely satisfactory manner as sAVAX operates on a similar calculated price feed from Chainlink and has a supply limit of 1M sAVAX or ~17.5 Mn $. Considering the ability to redeem liquid staked Matic combined with the fact that the proposed 15M liquid staked Matic supply cap is still within the supply cap limits for the likes of sAVAX which operates with a similar calculated price feed, a calculated price is well within the scope of the ask to increase the supply cap. We would like to further mention that due to the recursive nature of borrowing activity, the chances of insolvency are low from a market risk perspective.

This is a win-win situation for both parties as it helps unlock the power of liquid staking for the Polygon Defi ecosystem and ensures higher utilization of e-mode and migration of liquidity from Aave v2 → Aave v3.

[Vibin]

Just adding onto the point on calculated price feeds, Polygon staking already allows for redemptions with minimal unbonding period that makes the calculated price feed and DEX prices largely inline, as we have observed so far since MaticX going live in Apr '22.

Polygon has a 80 checkpoint (~2-3 days) unbonding period for unstaking. This means that users can get Matic back directly by unstaking MaticX with Stader in ~2-3 days. This makes the peg between calculated price (Calculated by Chainlink from Stader smart contracts, Matic/USD price) and DEX price very strong (in line with the argument for how Eth staking derivatives will likely need lesser LP liquidity post the enablement of withdrawals).

For instance, a 2% delta in price between the calculated price (based on the calculated price) and DEX price would imply a 240-360% APR (2% yield in 2-3 days → 240-360% annually) which will ensure the deviations of calculated price to spot price are unlikely to be meaningful.

Also, wanted to share an update on the pool migrations on Balancer, ~85% of our liquidity on Balancer has already migrated with the new pool already having ~15M$ in liquidity. We expect the remaining ~3M$ on the older pool on Balancer to also migrate over the course of this week.

[Kyros]

Good points put forward by @YJN58 in response to @MatthewGraham , followed by a great explanation of how things are supposed to play out .I still support the current proposal as it is laid out and think moving forward with it would prove to be very capital efficient for polygon Defi ecosystem and expand the money lego on polygon

[sandeepnailwal]

This is a great initiative to unlock the power of liquid staking of MATIC via LIDO and Stader Labs. Liquid staking will not only improve capital efficiency but will also pave the way towards further decentralisation of the network. This proposal is a definite win-win situation for both Polygon and Aave and would be great to see it implemented

[MatthewGraham]

Hi @YJN58,

Just so we are all on the same page here, I do support the general direction of the proposal. My concerns here are more focused on the need to provide more detailed risk analysis. Llama is actively working on a deeper dive into the on-chain liquidity.

There are ways beyond tweaking borrowing costs to entice Users into the recursive leverage trade on Aave. For instance, if both Lido and Stader transitioned Balancer liquidity to bb-a-wMATIC / Product pools, an estimated $17.1M (75%) of the $22.8M wMATIC liquidity on Balancer would be deposited into the Aave v3 wMATIC Reserve. There is currently $13.18M of wMATIC supplied to the Aave Polygon v3 and the linear pool would drive around 75% of the wMATIC supply on Balancer across the four pools, shown below, into the wMATIC on Aave. This represents a 129.8% increase in liquidity and will reduce the cost of borrowing wMATIC to sub 2.8% without changing the yield curve.

Just upgrading the Balancer Pools to bb-a-wMATIC / Product pools will reduce borrowing costs of wMATIC to sub 2.8% based upon the current amount of borrowing. These pools are expected to be ready in Q4 and are only pending front end upgrades. This integration by itself will enable the recursive trade to be profitable.

@YJN58

• Optimisation of the interest rate curve for both stMATIC and MATICx.
• Increasing the supply cap to 15M for both MATICx and stMATIC.

That said, if we want to entice more Users into the recursive trade. Only the wMATIC borrowing costs need to be changed. Changing the cost of borrowing stMATIC of MaticX does not make the recursive trade more profitable for Users. Increasing the LTV and Liquidation Threshold of stMatic and MaticX does.

The desired affect is to maximise the collateral’s ability to generate yield by borrowing as much wMATIC for as cheap as possible.

Something we need to be conscious of is what happens when large Liquidity Providers remove liquidity from DEXs and deposit it in Aave to perform the recursive yield trade ? We expect Liquidators to use USD to acquire wMATIC on market to pay back debt, receive stMATIC or MATICX and then sell the stMATIC or MaticX on market for a profit in USD terms. The liquidity is a significant variable. We need to be conscious of what could happen if whales more from providing Liquidity on Decentralised Exchanges to leverage Proof of Stake yield extraction. This Aave Liquidity Pool is not currently supported by gauntlet and in my opinion, it is better to lean on the side of caution. This is why I am asking for more insight into the liquidity. I also favour a staggering of the upgrades as well so we can monitor how each upgrade affects liquidity.

Something to keep in mind, post merge ETH Revenue per day is less now than what it was before the ETH Borrowing yield curve was changed. $5.3k/day compared to $3.5k/day now. This is an example of why I am suggesting a more data driven analysis rather than narrative.

Reference: https://dune.com/queries/566763/1061589

Below is a link to a file that shows that the leverage recursive yield trade can generate 26% return on investment with 6.29x Leverage (20 loops), 2% wMATIC borrowing costs, 1% slippage and no fees from an automated strategy provider (do it yourself style). The math is crude, it is not perfect, and serves to give a flavour of what is possible. The estimated yield of 26% exceeds what Liquidity Providers can earn on Balancer at the moment. Will liquidity providers shift from Decentralised Exchanges to earn yield on Aave…

Leverage MATIC Yield Trade.xlsx

I think we should do a deeper dive into the risks presented in this proposal and also separate each change into its own proposal. As we can see the discussion so far, many different changes are being discussed and it would be better if we approach them separately. This way one upgrade does not block another from proceeding.

I am particularly keen to learn more about the liquidity distribution and how the recursive trade affects what the whales do. The lowest risk path forward here for Aave is to await Balancer launches the Linear Pools inside a ConstantStablePool front end upgrade. This is expected to occur this quarter and materially change the wMATIC borrowing cost without needing to upgrade the Aave Liquidity Pools.

That said I think we are all heading in the same direction here. I am approaching this more cautiously than most based upon the comments to date.

[YJN58]

Hey @MatthewGraham

This was a typo:

I have corrected it to wMATIC.

[Crypto_wildwest]

Matic has been a proven asset and the way it has been performing in the market, there is no doubt on boosting it more on the AAVE market via this proposal. Love where it’s going.

[monet-supply]

I think adjusting MATIC rate curve is an easy win for Aave, similar to ETH rate curve changes on v2 which now see ETH users frequently earning 1%+ on their deposited collateral which helps offset their cost for borrowing stables or other assets.

Just a few comments from my side:

I don’t think adding stable rates is necessary or desirable. Leveraged staking users can be expected to always use the lower variable rate, and introducing stable rates creates some edge cases where it may be difficult for the market to reach a stable equilibrium.

It seems that adjusting wMATIC LTV / liquidation threshold was not proposed in OP. But if this is being considered, IMO it should be handled in a separate proposal because it is unrelated to this proposal’s goal of facilitating wMATIC borrowing and leveraged staking.

For rate curve parameters themselves, I think it would be preferable to set optimal borrow rate slightly above expected returns of liquid staking tokens. If they are set significantly below expected returns (eg. proposed 4.8% optimal rate vs 5.5% staking returns), it becomes more likely that utilization may exceed the kink and cause heightened interest rate volatility. This could potentially drive away long term staking users who could most benefit from this proposal.

IMO a more suitable rate curve would feature a optimal borrow rate of eg. 6% (slightly over expected yield), and then potentially a somewhat higher optimal utilization (eg. 75-80% vs proposed 70%). This will allow supply and demand for wMATIC to find equilibrium while still in the “flatter” part of the rate curve, avoiding unnecessary volatility while still preserving enough liquidity to ensure market safety.

[MarcZeller]

this proposal got refactored by @Llamaxyz and posted as a separated thread.

new proposal focus on wMATIC risk parameters update first, stMatic & MaticX updates can be done at a later stage on segregated proposals.

[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 eM…