Generalized LT/LTV Reduction on Aave

Summary

Reduce stablecoin LTs and LTVs across all markets.

Motivation

In response to the volatility risks highlighted by the UNI incident on Compound, this proposal aims to mitigate theoretical debt asset volatility by proposing a reduction in collateral asset borrowing power for stablecoins. While the UNI debacle was attributed to inadequate dual-bounded fallback oracle logic, a property not employed by Aave, leading to mispriced UNI debt and subsequent arbitrage activities, the observed upward volatility in longer-tailed assets suggests a prevailing β€œrisk-on” market sentiment. Therefore, we advocate for minimizing borrowing power concerning shorting these assets to prevent potential short squeezes and liquidation shortfalls. Subsequent recommendations will address borrow cap reductions for relevant long-tailed assets, which generate minimal revenue relative to the theoretical risks, especially in edge-case scenarios. Additionally, presently only 2.4%, 3.7% and 5.8% of supplied DAI, USDT and USDC on Ethereum are being utilized as collateral, indicating the generalized lack of demand (and thus revenue).

USDC, USDT and DAI (non-Avalanche/Metis, no DAI on Optimism), 80% β†’ 78% LT

Lowering USDC, USDT and DAI LT from 80% to 78% on said chains would induce $537k worth of liquidations, with a singular DAI-collateralized WETH and MKR debt position responsible for 93% of this value. It’s crucial to acknowledge the heightened market volatility at present, particularly regarding the growth of the underlying debt assets. Values have experienced significant fluctuations over the past few days. Therefore, we will allow users ample time to unwind risky positions before execution.

Chaos Labs Gnosis Risk Explorer

LUSD, sDAI and WXDAI, 80% β†’ 78% LT

Lowering LT from 80% to 78% on all available chains would result in just one ACI-affiliated position being liquidated on Gnosis.

Chaos Labs Arbitrum Risk Explorer

USDC.e (Arbitrum and Polygon) and USDbC (Base), 80% β†’ 78% LT

Chaos Labs Risk Explorer

USDC (86.25% β†’ 83% LT), USDT (80% β†’ 78% LT) and DAI (82% β†’ 80% LT) on Avalanche

Nearly all of the $21.2k in liquidations are attributed to USDC loopers with active positions held for over a year, of whom primarily engaged in farming WAVAX borrow incentives.

Chaos Labs Avalanche Risk Explorer

USDC.e (85% β†’ 83% LT) and DAI (83% β†’ 80% LT) (Optimism)

Chaos Labs Optimism Risk Explorer

Final Recommendations

Assets LT LTV Total Liquidated
USDC, USDT, DAI 80% β†’ 78% 77% β†’ 75% $537.4k
USDC.e, USDbC 80% β†’ 78% 77% β†’ 75% $9.5k
USDC (Avalanche) 86.25% β†’ 83% 82% β†’ 80% $21.2k
USDT (Avalanche) 80% β†’ 78% 77% β†’ 75% 0
DAI (Avalanche) 82% β†’ 80% 75% 0
LUSD, sDAI, xDAI 80% β†’ 78% 77% β†’ 75% $27.8k (ACI)
USDC.e (Optimism) 85% β†’ 83% 80% β†’ 78% 0
DAI (Optimism) 83% β†’ 80% 78% β†’ 75% $134
$596k
4 Likes

Hi Chaoslabs,

thank you for this post. Will the reduction happen in steps like 1% per week to be sure that these user can react and maybe the total amount liquidated will be lower?

And additionally, has there been any kind of simulation done by you regarding LT & LTV for the future.
What I mean is if we are settings those parameter higher in higher in an upcoming rising market, Aave will be stuck there more or less. We already see how these little changes liquidate nearly 600k. Imagine what is going to happen when Aave has 50bn TVL and LTV on assets as high as possible for best revenue. We won’t be able to lower them that easy. So we either have to liquidate user (not great) or live with the risk of bad debt.

What I want to say is that we have to think about this in the future and find the best solution somewhere in the middle.

2 Likes

It is important to have community concensus on what amount of forced liquidations is acceptable for these measures. $500k is not a small amount. @ChaosLabs please provide some additional analysis.

  • What are the assumptions on the how these changes provide further safety from the protocol?
  • How are you making this tradeoff between user liquidations and estimated future protocol safety?
    thank you.
1 Like

As user 0x6cd4471480e2969b3d696fbd17530e85112f3ff6 has been repaying the debt position, the reduction should no longer force it to go underwater. To further mitigate the impact on users and allow them to adjust their positions, we will implement some of the recommendations in two increments of 1%. The current total value liquidated following the first step of reductions is $9K.

USDC, USDT, and DAI (Excluding Avalanche/Metis, DAI on Optimism), 80% β†’ 79% LT:

image

USDC.e (Arbitrum and Polygon) and USDbC (Base), 80% β†’ 79% LT:

image

USDC (86.25% β†’ 85% LT) on Avalanche:

image

LUSD, sDAI and WXDAI, 80% β†’ 79% LT:

image

Step One Recommendations:

Assets LT LTV Total Liquidated
USDC, USDT, DAI 80% β†’ 79% 77% β†’ 76% $18k
USDC.e, USDbC 80% β†’ 79% 77% β†’ 76% $1.8k
USDC (Avalanche) 86.25% β†’ 85% 82% β†’ 80% $2.3k
USDT (Avalanche) 80% β†’ 78% 77% β†’ 75% 0
DAI (Avalanche) 82% β†’ 80% 75% 27
FRAX (Avalanche) 80% β†’ 78% 75% 0
LUSD, sDAI, xDAI 80% β†’ 79% 77% β†’ 76% 0
USDC.e (Optimism) 85% β†’ 84% 80% β†’ 78% 9
DAI (Optimism) 83% β†’ 80% 78% β†’ 75% $13

Thank you, @EzR3aL. We base our adjustments to Aave’s liquidation parameters and supply caps on simulations, putting the system through various stress scenarios. These include scaling up TVL, leverage, and volatility, which aligns with the concern you’ve raised. True, we’re always navigating the trade-off of maximizing the protocol’s revenue and mitigating the risk of bad debt, and we are aware of the difficulty of reducing the Liquidation Threshold if set too high. Our approach is to make gradual changes and constantly monitor the system’s state. This way, we can tweak the parameters in a timely manner to manage these challenges effectively.

5 Likes

We have updated the recommendations and total liquidated data in the table above to accommodate up-to-date conditions, and published a Snapshot for the community to vote on, starting in 24h.

We thank you in advance for your participation in the vote.

AIP-71 has been published for this proposal and is now live

Thank you in advance for your participation in the vote.

1 Like

As of today, 232 total accounts amounting to ~$19K are at risk of liquidation following the proposed changes.

The full list of accounts and their projected health score, given the updated LTs can be found in this sheet .

2 Likes

What percentage decrease in the E-Mode Category can be expected from this LTV Reduction?

Hi @jgusaino, this proposal does not impact the E-mode configuration.

1 Like

Thank you for the prompt response!

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