[ARFC] Increase Community Preference For Supply Cap Limits

[ARFC] Increase Community Preference For Supply Cap Limits

Author: ACI

Date: 2024-08-12


Summary

This proposal suggests increasing the informal supply cap restriction previously set by the community in the past vote [Temp Check] - Community Preference for Supply Cap Limits for LSTs.

Motivation

LST collateral and correlated asset borrowing with E-Mode is one of the primary use cases of Aave across networks. Previous discussions around supply cap methodologies resulted in the [TEMP CHECK] vote to limit the maximum supply cap of an LST to 75% of that LST’s circulating supply.

In the year and a half since that vote, Aave has continued to grow and new liquid restaking tokens (LRT) like weETH have shown strong demand as collateral in the protocol. LRTs have also seen strong proliferation to various chains with varying levels of the total supply spread across multiple L2s.

As a result, Aave is often the main location for storing LST and LRTs on any given chain and often represents an outsized share of the total supply on any given chain.

In this new environment, Aave has becomes the main location for LRT supply on chains like Scroll and Base, while also driving LST/LRT demand on chains like Polygon PoS. Simply put, when Aave increases LST supply caps, new supply comes to fill it.

LST and LRT have also had longer time in the market, with clear support for peg and asset profiles developing. Aave’s use of oracles based on the LST redemption rate and implementation of CAPO have allowed for reduced risk from the secondary market volatility.

This calls for a new limit on a collateral’s supply limit as a percentage of the total supply. This proposal suggests raising the community preference supply cap limit to 90% of the total supply.

Specification

  • The community preference will update from 75% of total supply to 90% of total supply.
  • Risk Parameters will be updated by Risk Service Providers based on the results of this ARFC as needed.

Useful Links

[Temp Check] - Community Preference for Supply Cap Limits for LSTs

Disclaimer

The ACI has not been compensated for the publication of this proposal.

Next Steps

  1. Publish a standard ARFC and collect community & service provider feedback before an ARFC snapshot.
  2. If the ARFC snapshot passes, this community preference becomes canon and risk service providers incorporate this into current and future recommendations.

Copyright

Copyright and related rights waived under CC0

4 Likes

Chaos Labs supports this proposal. We have observed that, for many LSTs and LRTs, leveraged yield farming on Aave rapidly becomes their top use case. As such, these markets can fill rapidly, requiring consistent cap increases that only slow as the cap approaches 75% of on-chain supply. We have also observed significant induced demand, in which on-chain supply increases following cap raises; this proposal will help further accelerate this process.

On the risk front, we find minimal increased risks associated with increasing from 75% to 90%. Assets for which we currently increase caps up to 75% are only the most robust, with a strong history of peg stability and mean reversion. However, the most important factors are the composition of assets borrowed against the asset in question, as well as the use of an exchange rate oracle. The latter prevents liquidation cascades, while the former is critical to assess the likelihood and size of potential liquidations. When increasing caps, we always ensure that positions with non-correlated borrows can be liquidated efficiently without incurring bad debt. This process will remain the same should this proposal pass.

5 Likes

The current proposal has been escalated to ARFC Snapshot.

Vote will start tomorrow, we encourage everyone to participate.

Summary

  • LlamaRisk supports the proposal to increase supply cap preferences to 90% for Liquid Staking Tokens (LSTs) and Liquid Reward Tokens (LRTs) across all Aave markets.
  • Market analysis:
    • The Base market is significantly affected, with the weETH supply cap nearing 95% of the total supply. Approximately 68% of the $82 million in borrows is concentrated in ETH-correlated assets, primarily due to extensive leverage looping between ETH and LSTs. A single borrower holds roughly 40% of the weETH supply.
    • In the Scroll market, the wstETH supply cap stands at 71%, with 80% of the $108 million in borrows focused on ETH-correlated assets. Of this amount, $72 million is in the ETH e-Mode, indicating a similar leverage looping strategy.
    • The Polygon market exhibits greater diversification, with wstETH at 70% of the total supply. Only 25% of borrows are in ETH-correlated assets, and ETH/LST leverage looping accounts for just 10% of the activity.
  • Liquidation risks are considered manageable due to using CAPO price oracles, which limit the impact on ETH-correlated borrow positions. Liquidators are incentivized to transfer liquidity from the Ethereum mainnet to L2 chains due to the potential for high liquidation bonuses and attractive LP yields in markets with thin liquidity.
  • Past liquidation events, potential discrepancies between supply and borrowing rates, and the ability to transfer liquidity across networks may introduce additional challenges. Each increase in the supply cap preference should be carefully evaluated based on the specific liquidity and market conditions.

LlamaRisk intends to conduct further research on liquidity requirements and liquidator response times in L2 markets to propose proactive measures to mitigate risks and support smaller Aave markets’ confident and sustainable growth.

Detail research below

Introduction

Given this proposal’s sensitive and risk-related nature, we aim to thoroughly assess the potential risks associated with increasing the cap preferences to 90% of the total supply for Liquid Staking Tokens (LSTs) and Liquid Reward Tokens (LRTs).

It is evident that on certain smaller chains where Aave markets operate, the Aave protocol has established itself as the leading platform. It attracts demand and, consequently, newly bridged assets as these supply caps are expanded. As a result, Aave is driving the growth of its new markets and contributing to the overall growth of asset supplies on these networks.

However, it is crucial to comprehend the circulation patterns of these assets, the nature of the borrow positions opened using such collateral, and the overall health of these positions. Furthermore, in the event of liquidations of an asset whose supply is heavily concentrated on Aave, the constrained on-chain liquidity necessitates robust bridging capabilities; hence, a thorough evaluation is required.

High supply assets & markets

To better understand the relevance of the supply cap problem, we have inspected current LST and LRT supply caps about their on-chain supply for different Aave markets. These are the findings:

  • weETH on Base, the supply cap is already at ~95% of all supply.
  • wstETH on Scroll market is at 71% of the supply.
  • wstETH on Polygon is at 70% of total supply
  • as for Ethereum market, weETH is at 50% of total supply.
  • All other LST and LRT assets on different markets are still far from the supply cap preference limit of 75%.

Therefore, this supply cap preference increase only concerns Base, Scroll, and Polygon Aave V3 markets. The assets of interest are weETH and wstETH.

Borrow positions

Scroll market

As for the Aave Scroll market, it is heavily focused on ETH-correlated assets and USDC. Consequently, 80% of the $108M borrows are in ETH-correlated assets.

image
Source: ChaosLabs Aave Dashboard

$72M worth of borrow positions are opened in ETH correlated e-Mode. Therefore, this market’s primary use case is ETH/LST leverage looping.

image
Source: ChaosLabs Aave Dashboard

Base market

Aave Base market possesses similar features, containing ETH-correlated assets as well as USDC and USDbC. Slightly less, 68% of $82M borrows are in ETH-correlated assets. Almost all of the ETH correlated borrow positions are in the corresponding e-Mode. Therefore, it contains the same ETH/LST leverage looping use case.

Moreover, the largest borrower, who also uses this market for the same use case, supplies more than half of the total weETH supply. It accounts for ~40% of the total weETH supply on the Base network.

image
Source: ChaosLabs Aave Dashboard

Polygon market

The Aave Polygon market contains more diverse asset types, where ETH-correlated assets only make up 25% of total supplies and borrow. Consequently, borrow positions are also more varied, with ETH correlated looping representing a mere 10% of the market’s activity.

image
Source: ChaosLabs Aave Dashboard

It cannot be concluded that ETH/LST leverage looping is the main use case for this market, and the wstETH supply on this market is slowly approaching the supply cap preference limit for other reasons.

Liquidation Risks

While many of the largest borrow positions are exposed to the looping and are of low health (due to e-Mode), CAPO price oracles limit the risk of liquidating ETH-correlated borrow positions.

Moreover, if the liquidations are bound to happen, the liquidators would be highly incentivized (due to the liquidation bonus) to bridge the LST/LRT liquidity from the Ethereum mainnet to these L2 chains and be able to perform the liquidations. Also, thin secondary market liquidity on these networks would offer high LP yields for such assets.

Nonetheless, it is important to dynamically monitor the overall health of the positions, as in addition to LST/LRT price deviations, a mismatch of supply and borrow rates for the looping users could also cause unexpected liquidations. It also needs to be considered that liquidators may not be able to instantly move liquidity from other, more liquid networks into the L2s.

The aggregated risk profile of the markets would not change drastically if the community preference for supply cap limits is lifted to 90% of the total LST/LRT supply. However, each supply cap raise requires an individual review covering specific liquidity and market conditions.

Next steps

LlamaRisk will further research this subject to evaluate the liquidity and liquidators’ needs in discussed L2 markets, understand how quickly the liquidity can be moved, and propose proactive measures to limit the discussed risks. This will help to scale the smaller Aave markets with higher confidence.

2 Likes

The current ARFC Snapshot has just ended, reaching both Quorum and YAE as winning option, with 516K votes.

Therefore, the ARFC to Increase Community Preference for Supply Cap Limits has passed, being considered canon from now on, updating community preference from 75% of total supply to 90%.

Summary

Following the proposal to increase supply cap preferences to 90% of the total chain supply for Liquid Staking Tokens (LSTs) and Liquid Re-staking Tokens (LRTs), LlamaRisk has conducted an analysis focusing on the Base and Scroll networks, where weETH and wstETH have high supply concentration ratios. The key findings are:

  • Aave’s supply caps drive the growth of LST assets on these networks, with Aave’s dominance temporarily increasing after raising supply caps but then typically decreasing as LST supply moves to other protocols.
  • Sufficient on-chain liquidity is crucial for initiating liquidations and capitalizing on liquidation discounts. Current Base and Scroll liquidity is sufficient based on at-risk collateral and past liquidation data.
  • To be more conservative, DEX pools should be monitored to ensure the supply cap does not exceed 2x the liquidity available within the liquidation bonus slippage.
  • Proactive monitoring and measures are recommended, including:
    • Ensuring sufficient liquidity before raising supply caps on heavily concentrated LST assets, as estimated via the liquidity target formula.
    • Monitoring liquidity-to-supply ratios after aggressive supply cap increases to ensure stability.
    • Incentivizing liquidity bridging on the L2 network when liquidity falls below targets or adjusting supply caps and other parameters to make the LST collateral less attractive.
    • Maintaining Aave’s dominance below critical levels (>0.95) to ensure sustainable network growth.

Future research will focus on assessing the feasibility and impact of cross-chain liquidity on liquidations, developing agent-based simulations to model MEV actors, and investigating the implications of relying on cross-chain liquidity for L2 liquidations.

Expand to see full research

Introduction

This note follows the Aave community proposal that increased the supply cap preferences to 90% of the total chain supply for Liquid Staking Tokens (LSTs) and Liquid Re-staking Tokens (LRTs). LlamaRisk has provided an initial note on the governance forum, covering an overview of the risks associated with this proposal.

Although the risk profile was considered to change insignificantly in the face of more drastic supply caps, LlamaRisk aimed to further research liquidity needs and liquidator response times in L2 markets, proposing proactive measures to mitigate risks and support smaller Aave markets’ confident and sustainable scaling.

This second note focuses on the two Aave markets with the highest LST supply ratios: Base and Scroll, where:

  1. weETH supply on Base is at ~84% of all supply
  2. wstETH supply on the Scroll market is at 71% of the supply

Dominance of Supply

Aave’s Influence on In-Network Supply

As initially suggested, Aave attracts demand and newly bridged assets as supply caps on these smaller markets are increased. Consequently, Aave drives the growth of its new markets and contributes to the overall asset supplies on these networks.

On-chain data confirms this claim, indicating a clear correlation between LST supply caps on Aave and overall LST supply on Base and Scroll.

image

Source: Dune

image
Source: Dune

Therefore, Aave’s proposal to enable LST leverage looping provides a clear incentive for users to bridge LST assets to these networks.

Aave’s Influence on Overall Network Activity

In addition to Aave’s dominance, we are also interested in the consequences of increasing the supply cap. Specifically, we aim to understand whether bridged supply remains solely in Aave or is distributed to other protocols, increasing on-chain liquidity without further amplifying Aave’s market dominance.

The following graphs for Base and Scroll chains indicate that Aave’s dominance increases immediately after a supply cap increase but starts decreasing within a few days, resulting in a larger TVL of LSTs in other protocols.

image
Source: Dune

While on Base, the supply ratio always retracts after sharp increases, on Scroll wstETH, supply ratio reduction is currently slower than historically, decreasing by 2% after the last hike and a peak of 89% one week ago.

image
Source: Dune

These findings suggest that such supply cap increases do not lead to a persistent rise in Aave’s dominance but instead support the long-term growth of these networks by encouraging the distribution of supply to other protocols. This data further supports the rationale behind aggressive supply cap adjustments.

Liquidity

The high concentration of LST supply on Aave necessitates sufficient liquidity for both borrowed and collateral assets to enable users to perform liquidations efficiently. More precisely, a borrowed asset is needed to initiate the liquidation, and the collateral asset of a liquidatable debt position is received at a discount. From the liquidator’s perspective, collateral assets must be exchanged (or redeemed for ETH in the case of LSTs) to capitalize on the discount and profitably finalize the liquidation.

Liquidity Need for Liquidations

Let’s illustrate a classical borrow position example when LST leverage looping is performed: a debt position on Base where 700 WETH are borrowed using 900 weETH as collateral. Suppose this position needs to be fully liquidated. If the liquidator does not directly possess the required 700 WETH, it can be acquired in the following ways:

  1. Using DEX liquidity on Base to swap other assets for WETH
  2. Bridging WETH from the mainnet
  3. Taking a flashloan in WETH

Then, WETH is used to repay the debt and, in turn, receive discounted weETH collateral. After receiving weETH, the liquidator needs to swap weETH for WETH (or redeem it for ETH) to capitalize on the discount. This can be done by:

  1. Using weETH/WETH DEX liquidity on Base
  2. Bridging weETH to the mainnet and exchanging weETH there

However, if the WETH was acquired using flashloans, the only possibility is to use DEX liquidity on Base. Moreover, bridging weETH to the mainnet may take significant time, exposing the liquidator to the risk of losing the discount during volatile market conditions.

Therefore, this example demonstrates that having adequate available on-chain (L2) liquidity is highly preferable and crucial for efficiently executing liquidations.

Liquidity on Base

DEX liquidity for weETH on the Base network is currently $7.8M, with a weETH share of $3.5M. All liquidity pools are weETH/WETH, which is a suitable pair for the above-described LST liquidations scenario.

image

Source: DefiLlama, August 29th, 2024

This liquidity represents 14% of the total weETH supply on the Aave Base market.

Nonetheless, the exchange rate for an asset is non-linear concerning the amount that would be sold due to slippage. Therefore, evaluating the liquidity situation by inspecting the following metric: 2x liquidity available within liquidation bonus as price impact is also important.

At the time of writing, 480 weETH (~$1.15M) could be swapped to WETH with a 7.5% liquidation bonus. It represents ~2.3% of the total weETH supply on Aave’s Base market.

image
Source: Odos aggregator, August 29th, 2024

Liquidity on Scroll

Liquidity on Scroll for wstETH is concentrated on Nuri DEX, where a wstETH/WETH liquidity pool with a TVL of $4M and a wstETH share of $0.3M is available, and on Ambient DEX where a wstETH/ETH liquidity pool exists with a TVL of $7.31M and wstETH share of $2.91M.

image
Source: Nuri exchange, 29th of August, 2024

image
Source: Ambient exchange, 29th of August, 2024

This liquidity represents 19.2% of the total wstETH supply on the Aave Scroll market.

3,000 wstETH ($7.9M) could be swapped to WETH with a 7.5% liquidation bonus. It represents ~13.4% of the total wstETH supply on Aave’s Scroll market.

image
Source: Odos aggregator, August 29th, 2024

Positions at Risk

Multiple risk factors must be considered to understand the required liquidity. Additionally, historical liquidations and current positions at risk provide a conservative estimation of LST liquidity needs.

Low Health Positions

Currently, $1.82M of wstETH collateral is at risk on Scroll, which amounts to 45% of the total wstETH liquidity on the network.

image
wstETH collateral at risk on Aave Scroll, Source: ChaosLabs dashboard, August 29th, 2024

In contrast, a much less tense situation can be observed with weETH collateral. Less than $50K of weETH collateral is at risk on the Base market.

image
weETH collateral at risk on Aave Base, Source: ChaosLabs dashboard, 29th of August, 2024

These differences indicate that the situation can vary, and needed liquidity thresholds can change drastically over time and across different markets.

Past Liquidations

Given that L2 markets are less mature, there is no extensive historical data to monitor the extent of liquidations during high-market stress periods. Nonetheless, available data can provide floor estimates of the required liquidity.

On Base, the largest one-day weETH collateralized loan liquidations were $400K during recent market stress on August 5th, 2024.

image
Source: Dune

On Scroll, wstETH liquidations were almost non-existent, with maximal one-day liquidations reaching only $20K.

image
Source: Dune

While both markets indicate no stress induced by liquidations, the data cannot conclude that liquidity needs are completely satisfied.

Other Risk Factors

In addition to collateral at risk and historical liquidations data, and more importantly, since our focus assets are LSTs, other risk factors are also significant.

Upward market price manipulations are mitigated by Aave’s CAPO mechanism that is used for all LSTs on different Aave markets; therefore, a manipulation risk is only present for a downward peg of an LST asset.

Firstly, the risk of slashing is always apparent when using LSTs as collateral. While such events are unpredictable, they would cause position liquidations on all Aave markets, making LST-to-base token liquidity scarce even on Mainnet, discounting the ability to bridge assets and perform liquidations efficiently.

Secondly, LST-to-base de-pegs may disturb the market, especially if the main use case is LST leverage looping. As this is the case for Base and Scroll, elevated caution is necessary when estimating liquidity needs.

Historically, such risks were minimal; therefore, only the general collateral at risk and the sizes of LST loan positions must be monitored to evaluate these changing risk factors.

Estimated Liquidity Coverage

Based on the analysis of positions at risk, historical liquidations, and other risk factors, it is possible to dynamically estimate the liquidity needs for each high-supply LST asset on Aave L2 markets.

The relative liquidity of an LST asset for which Aave increases the supply cap to >75% of the total on-chain supply should also grow. In that case, the liquidity-to-supply ratio should be stable or increase. This would ensure that the overall LST asset supply grows more organically and that the Aave market does not become increasingly exposed to the risk of insufficient liquidity for liquidations.

The collateral at risk metric would need to be monitored to quantify the liquidity needs and the target liquidity re-evaluated flexibly. The estimate is as follows:

image

The current liquidity of both weETH on Base and wstETH on Scroll remains above the target; therefore, the native chain liquidity can be deemed sufficient for both assets. However, the liquidity-to-supply ratio must be monitored after each aggressive supply cap increase.

Recommended Actions

An extensive analysis of LST liquidity needs has not indicated elevated risks on the markets and assets of interest. However, given the dynamic nature of these supply cap increases, we recommend continuously monitoring the situation and facilitating cap increases by suggesting the following actions:

  • Before raising the supply caps on an LST asset that is already heavily concentrated, sufficient liquidity needs to be ensured as estimated via the liquidity target formula.
  • After raising the supply caps, the liquidity-to-supply ratio needs to stabilize in the days following the supply cap increase.
  • If the liquidity of a monitored high-supply asset falls below the targets, 1) bridging the liquidity needs to be incentivized on the L2 network (to be taken care of by the Aave Liquidity Committee), or 2) supply caps need to be adjusted downwards (if the supply cap has not been reached yet), and other parameters need to be adjusted to make the LST collateral less attractive.
  • Aave’s dominance should be at a manageable level (>0.95) for a sustained period. Otherwise, it would mean that the effective growth of the network is sustainable.
1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.