[ARFC] MaticX Polygon v3 Upgrade

Governance
1

title: MaticX Polygon v3 Upgrade
author: @Llamaxyz - Defi_Consulting(@MatthewGraham) @scottincrypto
created: 2023.02.02

Simple Summary

This ARFC proposes updating the MaticX Reserve on Polygon to facilitate an expansion of the MaticX/wMATIC loop strategy in the anticipation of rewards being introduced in the near future.

Abstract

The utilisation of the MaticX reserve has been trending higher over time and with the introduction of deposit rewards, there is an expectation the SupplyCap will be reached prohibiting additional MaticX from being deposited.

Increasing the SupplyCap enables the recursive MaticX/wMATIC strategy to grow beyond the current constraints. Enabling borrowing with a reasonable BorrowCap and the introduction of borrowing rewards will lead to significant utilisation of the MaticX reserve.

This proposal facilitates the expansion of MaticX yield-based strategies on Aave v3 Polygon by creating more favourable bootstrapping conditions. The introduction of liquidity mining rewards are expected to drive notable inflows and usage to the v3 deployment.

Motivation

Over the previous months, @Llamaxyz has been working with various communities to craft favourable conditions on Aave v3 Polygon to facilitate the creation of several yield aggregation products. The below proposals details are applicable:

This proposal is a continuation of the above work.

In support of the upcoming liquidity rewards distribution, this proposal seeks to increase the SupplyCap, enable borrowing and introduce a BorrowCap sufficient to create a notable deposit yield.

The current SupplyCap is set at 6.0M units of MaticX and has not been updated since MaticX was listed on Aave v3. When Llama listed MaticX, the risk parameters were set very conservatively. With improved liquidity conditions now and soon even better conditions when the new MaticX / bb-a-wMATIC pool launches in February, this SupplyCap can be increased.

With the introduction of SD rewards, it is fair to conclude, based upon past experience, there will be a material inflow of deposits into the MaticX reserve. With utilisation already at 66.23% and deposits climbing in both USD and units of MaticX terms, as shown below, it is highly expected that the SupplyCap will indeed be reached with the introduction of SD rewards.

The charts below show the historical volume of MaticX and USD value of MaticX deposited in the reserve over time.

The recent push to launch automated strategies has led to material improvements in utilisation of the SupplyCap. With more strategies to be launched in the near future, the SupplyCap will become a burden preventing other communities’ products from growing TVL. This problem will be exacerbated with the introduction of rewards. If the SupplyCap is reached, then no new MaticX deposits can be made to improve the health factor of these strategies. This causes the strategies to deleverage and generate a lower return for their users. This suboptimal outcome discourages builders and creates a need to proactively manage risk parameters and work closely with communities that build on top of Aave. The chart below shows the distribution of users’ funds in the pool.

The chart below shows the distribution of MaticX deposits.

The chart below shows that 92.3% of all debt drawn against MaticX collateral positions is wMATIC. This highlights the significance of the MaticX and wMATIC relationship.

The chart below shows the amount of MaticX that would be for a given price. If the price of MaticX falls to 35 cents, 2.05M units of MaticX will be liquidated. This is shown as the 3.45% allocation in the above pie chart which represents the largest MaticX collateral and non wMATIC debt position.

By aggregating all positions together, we can simulate the volume of MaticX deposits liquidated for a predefined percentage fall in the MaticX price. The chart below shows the volume of MaticX deposits liquidated for a 50% fall in the price of MaticX. A 50% price fall in MaticX would trigger 1.7M units of MaticX collateral positions being liquidated as of the 27th January 2023.

On-chain liquidity for MaticX, measured in units of MaticX, has been steadily sideways to slightly higher overall during 2023.

11

The chart below shows the USD value held within the main liquidity pools across Polygon.

The MaticX oracle is a calculated oracle and does not rely on spot pricing which removes the ability to liquidate users on Aave through manipulating MaticX spot pricing. The MaticX/USD Chainlink Oracle was used upon listing MaticX and wMATIC/USD when listing wMATIC. The MaticX oracle utilises the MATIC oracle in calculating its feed price. It is worth noting that to influence the spot price of MaticX through repeat entries on a lending market, the attacker would need to offset the flow of arbitrage traders profiting from the price variance. ie: If MaticX spot price is above the calculated feed, then swap MaticX for wMATIC and stake, and if MaticX’s spot price is trading at a discount, swap wMATIC fro MaticX, deposit Aave, borrow wMATIC and swap to MaticX in a recursive loop. Some arbitrage flow would be limited by the bridge transaction times and the Stader provided deposit liquidity.

When Borrowing is enabled, it becomes possible to arbitrage the spot and the calculated oracle. The arbitrage loops makes it more expensive to move the MaticX away from the Net Asset Value as it becomes profitable to run a counter strategy. The combination of having a base asset, wMATIC, with a staking contract makes this feasible and Aave’s liquidity would then make it easier to execute the arbitrage strategy. The chart below shows the DEX Aggregator price relative to the Calculated price feed. It shows the spot pricing closely tracking the calculated feed.

This is supportive of enabling borrowing of MaticX on Polygon Aave v3.

Looking at just the main liquidity pool, a 2% price impact swap will facilitate a swap size of 5.13M units. For 5% price impact, 6.12M of MaticX can be swapped. The entire SupplyCap on Aave could be purchased for less than a 5% price impact from a single DEX pool.

Expanding beyond the main liquidity pool and routing a swap via 1inch on Polygon, the below swap sizes can occur for a predefined 2% price impact. This is optimised for speed when using the 1inch’s new Fusion upgrade. There is also additional liquidity available by Stader, which is not likely to be integrated into the 1inch platform.

Screenshot 2023-02-02 at 17.29.30
Screenshot 2023-02-02 at 17.29.301200Ă—570 91.9 KB

With reference to the MaticX collateral and now wMATIC debt positions, the below chart shows the amount of price impact the main MaticX/wMATIC Balancer pool would occur when all users positions are liquidated in the event of a 50% price fall event. Liquidating all MaticX positions due to a 50% fall in the price, would cause -0.33% price impact if swapped at once. Due to improving liquidity over time, the price impact of liquidating users due to a 50% fall in price is reducing. This may change with the introduction of more structured products that utilise uncorrelated assets as the initial collateral deposit. However, these strategies can also be automated and unwind the leverage gradually to avoid liquidation.

It is worth noting here, a new MaticX / bb-a-wMATIC pool will be deployed (February) that will have an amplification factor of 500 and will replace the current MaticX/wMATIC pool with an amplification of 50. This 10x amplification factor increase will significantly improve liquidity measured in swap size for a given price impact through the main Balancer Pool. Migrations of liquidity from one pool to the next occurs fairly quickly with 86% of the liquidity migrating within 12 days last time the pool was migrated.

Having worked with the Balancer team we are confident the vast majority of the liquidity will be migrated to the new pool quickly. This is largely due to the end user preference to use automated strategy platforms like Beefy, InstadApp, Cian, and others…

With Borrowing Enabled and borrowing rewards introduced, the recursive deposit and borrow the same asset loop, is expected to inflate deposits and borrowings in the reserve. An ideal outcome would be for on-chain liquidity to benefit from the introduction of borrowing rewards. However, in the past users have shown a preference to enter a single asset recursive loop. This strategy would be limited by the lesser of the LTV, SupplyCap and BorrowCap parameters.

Due to the imminent SD rewards emissions, we are opting for a more generous BorrowCap allowance than otherwise would have been suggested. A BorrowCap of 5.2M which is still less than the (Uoptimal + 10%) x SupplyCap methodology indicates, but sufficient to enable the Aave community to gauge the market’s appetite for borrowing MaticX. Future proposals can revise this parameter. It is worth noting that during discussions with some yield strategy managers, it was highlighted that some vaults are optimised for yield across various strategies which include providing on-chain liquidity.

Since listing MaticX, the Loan To Value (LTV) and Liquidation Threshold (LT) has not been revised. This proposal recommends revising the LTV and LT from 50% to 58% and 65% to 67% respectively. These parameter adjustments will not affect the recursive loop strategies MaticX/wMATIC that utilise the Matic eMode category. These adjustments will encourage users to deposit MaticX as collateral and borrow assets other than wMATIC.

The chart below shows the MaticX price and how the volatility of MaticX has varied overtime. This is reflective of the underlying wMATIC assets volatility.

The chart below shows Annualised Volatility for various time sample periods.

The above data indicates a strong case for increasing the LTV for MaticX and wMATIC. As this proposal focuses on MaticX, any future wMATIC upgrades will be presented via a separate proposal.

Increasing the LTV and LT as defined above may be just enough, but probably not, to support building leverage long and short products with ~2x leverage which also minimise carry costs through utilising the MaticX/wMATIC recursive loop. Some communities are exploring the feasibility of this new type of structured product. At 65% LTV, ~2x leverage is possible as proven by Index Coop’s MATIC2x-FLI-P that uses Aave v2 and wMATIC as collateral. However, it is not yet known if the recursive loop with 58% LTV can enable free carry leverage trading products to be developed on top of Aave v3. To be clear, this does not influence the risk parameters setting and is intended only to provide context around some of the benefits associated with higher LTV and LT parameters.

Interest Rate

With Borrowing enabled, the following rationale was applied when developing the MaticX interest rate.

For the recursive strategy, it is highly desirable the deposit APR changes gradually with varying Utilisation. As the deposit rate for MaticX offsets the borrowing costs of wMATIC, any heightened volatility in the deposit rate of MaticX leads to more frequent rebalancing. This is more prominent within strategies trying to maximise yield. Due to the recursive loop, the end user experiences amplified volatility of the delta between the MaticX deposit rate and wMATIC borrowing rate.

To reduce this volatility, the Slope1 parameter should be minimised. A higher Uoptimal also reduces the volatility of borrowing costs by acting to distribute the increase in the borrow rate over a greater Utilisation horizon. An initial MaticX Slope1 parameter of 4% is proposed.

In accommodating a low Slope1 parameter and achieving a suitable Borrowing rate, a Base can be introduced. The borrow rate at the Uoptimal utilisation point is equal to the sum of the Base and Slope1 parameter. An initial Base parameter of 0.25% is proposed. The borrow rate at the Uptimal point is 4.25%.

An initial Uoptimal parameter of 65% is proposed. In time this can be increased depending upon the size of the larger borrowers of MaticX holdings. It is worth noting that swapping 5.125M of wMATIC to 4.79M of MaticX incurs a 2% slippage and the liquidator receives a 10% liquidation bonus payment which more than offsets the price impact of a swap. This compares favourably with the initial BorrowCap of 5.2M units of MaticX.

An initial Slope2 parameter of 150% is proposed and is support by Gauntlet.

Screenshot 2023-02-14 at 13.37.12
Screenshot 2023-02-14 at 13.37.121182Ă—726 45.5 KB

Specification

The following risk parameters are being proposed and have been reviewed by Llama.

Parameter Proposed Value
Loan To Value 58.00%
Liquidation Threshold 67.00%
Liquidation Bonus 10.00%
Liquidation Protocol Fee 10.00%
SupplyCap 8.6M units
Borrowing Enabled
BorrowingCap 5.2M

The following interest rate parameters have been reviewed by Llama.

Parameter Proposed Value
Base 0.25%
Uoptimal 45.00%
Slope1 4.00%
Slope2 150.00%
Reserve Factor 20.00%
Stable Borrowing Disabled
Stable Slope1 0.50%
Stable Slope2 150.00%
Base Stable Rate Offset 2.00%
Stable Rate Excess Offset 5.00%
Optimal Stable To Total Debt Ratio 20.00%

Next Step

2-3 weeks of discussion on the forum and then a Snapshot vote will be held. If the vote goes well, then Llama will prepare the payload for BGD to review soon thereafter.

Copyright

Copyright and related rights waived via CC0.

6 Likes
2

The proposal aims primarily to facilitate an expansion of the MaticX/WMATIC loop strategy. Generally, we see the value in the proposal and think that, like other loop strategies, the protocol can benefit from it.

There are several areas in this proposal leaning heavier towards the rewards side of the risk-reward ratio, for which we would like to recommend a more prudent and phased approach.

Supply Cap

We generally support increasing the supply cap and providing a low Slope1 on the IR curve to facilitate MaticX/WMATIC loop strategy. However, the proposed increase is too steep at the moment. We believe that a supply cap of 16M, which represents 37% of the total market cap of the token, creates high exposure to MaticX.

Regarding the new Balancer pool, while it may improve liquidity, we recommend waiting for the pool deployment before taking it into account in the analysis. As a general guideline, we should avoid theoretical liquidity and account strictly for what is available on-chain.

With the current utilization of the supply cap at 67%, we recommend an initial increase of 30% to 7.8M. This will allow room to absorb new supply while analyzing the user behavior following the enablement of borrows. Relaxations can be made over time, but only after more liquidity is present, demand is exhibited, and usage patterns become more evident.

Borrow Cap

The use case for arbitrage through borrowing is clear. However, a borrow cap of 6M MaticX, representing over 14% of the total market cap of tokens, seems excessive. Therefore, we propose starting with a lower cap to verify that there is no risky use of borrows and evaluate an increase over time.

We recommend an initial borrow cap of 90K, which is 10% of the highest daily volume ever for MaticX and above p90 of daily volume.

LT/LTV

We recommend leaving the current LT/LTV configuration without change. We do not think the LTV should be increased to that of MATIC (65%) as MaticX, by definition, encapsulates a volatility risk greater than MATIC’s.
The increase in LT and LTV does not promote the loop strategy but instead creates overexposure of the protocol to MaticX, which, together with the liquidity considerations described above, creates a potentially dangerous scenario of bad debt. Increasing these parameters, as mentioned before, should be considered only after borrow usage patterns become more evident.

2 Likes
3

Hi @ChaosLabs,

At Defi_Consulting, we are wondering if Choas Labs can elaborate on why the proposed Supply Cap and Borrow Cap methodology is not being applied here. We kindly suggest revising the Supply Cap and Borrow Cap methodology publication to incorporate the rational used above or alternatively, aligning the recommendation here with the proposed methodology.

Would it be possible for Chaos to elaborate on how was the 30% and 90k units figure was calculated ?

Regarding the daily trading volume, the Borrow Cap proposed, 10% daily trading vol, aligns with 3 of 4 highest days of trading volume on the Balancer pool.

image

Given an LST is primarily for holding, liquidity would have been a more applicable measure relative to trading volume. It is worth noting, a Borrow Cap set at 90k is not likely to encourage rewards being distributed to users borrowing MaticX. Even with deposit incentives, the recursive loop within just the MaticX reserve becomes more challenging to execute.

Given Aave’s past history with deposit and borrow rewards, is there room to increase the Borrow Cap ?

0.5% price impact on the single largest DEX pool is still conservative given the 10% Liquidation Bonus being proposed.

image

4

As stated in our previous response, because this proposal is designed to introduce new borrow activity, we want to observe and analyze it before setting more aggressive parameters. This analysis includes a variety of usage patterns, such as the size of positions, borrowed assets against MaticX, and emode usage, to name a few.

As a guideline, we view supply and borrow caps as dynamic parameters and do not recommend increasing caps to more than double the current supply in a given proposal. This was the motivation behind the 30% increase from the current utilization of 67% (67%*2 = 134%, we rounded it to 130%). For the borrow caps, we explained our considerations in the analysis in our previous comment.

We agree that the initial cap is conservative, but this aligns with our approach to allow for initial flow and observe behavioral patterns before setting the long-term optimal parameters. In addition, we will be actively monitoring the utilization of the caps and recommending updates to allow for healthy growth alongside our proactive risk management.

5

Hello,

Please remember that unstaking is live & working on the polygon blockchain and anyone can mint/burn MaticX for the fair underlying value in roughly 3 days.

Any peg variation are heavily arb and there’s no particular reason to consider maticX (or stMatic) as illiquid.

Worse case scenarios would present temporary excess debt that could be arb in a few days.

With ACI we’re supportive of much less conservative caps, LSD/Matic is key strategic assets for Aave Growth.

let’s try to find a consensus path above current @ChaosLabs proposal and more conservative than @Llamaxyz proposal to “allow for initial flow and observe behavioral patterns before setting the long-term optimal parameters” while not slowly growth too much.

1 Like
6

Below are Gauntlet’s Recommendations:

LT/LTV

LT → 67%

LTV → 58%

image

The above is a 15-minute MATICX/MATIC trajectory over the past few months. We note the maximum dislocation was roughly 1.5% during early November. The positive drift of MATICX/MATIC can be attributed to MATICX gradually accruing yield over time.

Moreover, unlike wstETH, MATICX can be directly redeemed into MATIC via smart contract, in addition to the market. The published exchange rate via StaderLabs is 1.0515, while the market rate is 1.0503, which reflects the 90-checkpoint delay associated with unstaking MATIC.

We propose increasing the LT from 65% to 67% and the LTV from 50% to 58% to better capture its risk profile and bring it more in line with other staking derivatives in relation to their underlying (i.e., wstETH has 79.5% LT, WETH has 82.5% LT on v3 Ethereum) while taking into account its inherent risk as a derivative of MATIC.

Supply/Borrow Cap

Supply cap → 8.6M

Borrow cap → 5.2M

Per our supply/borrow cap methodology, Gauntlet analyzes a number of metrics to holistically capture the risk profile of an asset in order to create caps that both minimize tail risk and encourage protocol growth.

Our conservative recommendation for the supply cap for MaticX, after considering its price manipulation potential, DEX liquidity, and its circulating supply, comes to 5.2M. Our aggressive recommendation comes to 8.6M. It is up to the community to align on risk preferences, as we note in this thread here. Please note that aggressive recommendations align with higher risk tolerance as we discuss in that thread.

Given the current supply of 4.1M and the current cap of 6M, we recommend that any supply cap increases for now are upper bounded by our aggressive recommendation of 8.6M. As Chaos mentions, it is not prudent to draw upon anticipated supply and liquidity to formulate current parameters.

Our conservative recommendation for the borrow cap comes out to 5.2M. Given that borrowing has not been enabled yet, 5.2M will be an initialization cap that mitigates tail risks while giving ample exploration opportunity for MATICX borrowing usage.

Interest Rate

The proposed IR curve is inconsistent with the risk profile of MATICX relative to MATIC and wstETH - in particular, the proposed Uoptimal as well as Slope 2. Given that MATICX is less established (lower volumes, lower market cap, lower circulating supply) than wstETH/stETH, Gauntlet’s analysis shows that the proposed IR curve is too aggressive compared to the curve for wstETH.

Gauntlet recommends Uoptimal = 45% and Slope2 = 150% to better capture the risk profile for MATICX in relation to wstETH.

In addition, Gauntlet has analyzed the conversion rate of MaticX. The conversion rate is determined by totalShares / totalPooledMatic, which change together during the mint and withdraw functions via the sendMessageToChild function. This seems to be the only place where these numbers are adjusted.

totalShares refers to the total amount of MaticX in existence, which only seems to change during burn/mints

totalPooledMatic sums all the staked Matic, so it seems like users need to stake Matic to affect this number.

There does not seem to be a capability for the contract to examine how much MaticX it has and add to the totalSupply. Moreover, any staking call is associated with a mint call (submit function), so users can’t stake without minting MaticX.

7

Hi @Pauljlei,

Thank you for the feedback. We have updated the forum post and created a Snapshot for this proposal.

https://snapshot.org/#/aave.eth/proposal/0x8abe6a2ae1134d6886f460d6648dc3a5d2f789e1b94e78f496e7f5ec5ff38697

Thank you to everyone who participates in the vote.

8

Hi Everyone :wave:

The SupplyCap on MaticX was reached yesterday and has just dropped back to near 95% at the time of posting here. As a result this proposal is needed urgently to enable Aave to grow MaticX TVL and wMATIC revenue from the looping yield strategy.

Llama has already started working on this proposal and will prioritise it to maximise the benefits from the SD rewards being distributed by Stader Labs.

Screenshot 2023-02-18 at 11.41.47
Screenshot 2023-02-18 at 11.41.471952Ă—866 83.3 KB

1 Like
9

By employing Gauntlet’s supply and borrow cap methodology, our conservative supply cap calculation is 5,769,000, and our aggressive supply cap calculation is 9,615,000. Thus, if the community has an aggressive risk preference, they may choose to increase the cap to 9.615M

image (1)

1 Like
10

Hi Everyone :wave:

An update on this proposal, it has completed the internal review at Llama and is now with @bgdlabs for peer review. We are intending to submit the proposal for voting later this week.

11

Hi Everyone :wave:

This proposal is now live for voting.

Quorum is not yet been reached and will finish over the weekend.

1 Like