Mortgage buffers tighten, Bitcoin loans struggle to withstand a 4.7% decline

The governance process for the decentralized lending protocol Aave is advancing a new proposal that, if passed, would give users who use bitcoin-like assets as collateral higher borrowing limits but correspondingly compressed liquidation buffer space.

Borrowing rate to rise from 73% to 81%

The proposal states that the maximum loan-to-value ratio (LTV) of WBTC and cbBTC mortgages in Aave V3's Ethereum Core Pool would be increased from 73% to 81% while liquidation threshold would be raised from 78% to 85%. This means for every $1 of Bitcoin collateral, the maximum amount borrowed will increase from approximately $0.73 to about $0.81.

LlamaRisk announced on September 21 that the proposal had entered the Snapshot voting phase, with polls to open within 24 hours. As of press time, results and follow-up implementation have not been confirmed; the above parameters are still in a proposed state and not set on an effective chain.

It is worth noting that the nominal gap between LTV and liquidation depreciation will narrow from 5 percentage points to 4 percent. The price buffer for users touching the settlement line from their highest borrowed state, which would be reduced from about 6.4% to about 4.7% by converting it into a collateral-price fluctuation assuming unchanged debt with Bitcoin prices falling separately.

Adjustments beyond the ethereum core network

This parameter adjustment is not limited to the Ethereum Core Pool. The proposal also plans to increase WBTC's regular LTV on Arbitrum by 5 percentage points, and Base Network cbBTC by 8 percentage points; WETH, wstETH and weETH in the Ethereum core pool will each receive 0.5 percentage point of LTV upgrades.

However, these figures only describe the maximum volume of collateral per unit and do not represent how much debt users will actually add. The proposal does not disclose full position data including mortgage balances for each impacted asset, corresponding debt sizes or account health distribution so that overall credit impact cannot be quantified.

Discrepancy between settlement speed data and extreme market models

LlamaRisk's research covers the settlement performance of three chains: Ethereum Core, Arbitrum and Base from August 2025 to August 2026. The data shows that there were a total of 7,206 ETH settlements involving about $618 million; 2,621 BTC settlements involved about $358 million.

In a separate test in two stress periods, February and October 2025, 100% of the liquidated positions were disposed within 5 minutes after triggering clearing profit across all listed markets. No bad debt was generated in February; about $390,000 of event-level bad debts occurred in October (equivalent to roughly $128 million), but LlamaRisk said they are unrelated to ETH and BTC collateral covered by this proposal.

The model then combines the hourly price fluctuations with clearing penalties for each reserve pool to derive an upper limit on the liquidation thresholds. Of these, 99.9% are adverse one-hour movements; ETH is 11.85%, and BTC about 5%. But the same dataset also records a more extreme scenario: ETH fell by 24.27% in less than an hour, while BTC was 10.72% (written as 11.15% in the proposal summary) and internal inconsistencies existed.

It remains to be seen whether historical performance will cover future risks.

This gap constitutes residual risk. LlamaRisk's framework assumes that the predictor is updated normally and the clearing robot responds in a timely manner, with maximum leverage positions not remaining in extreme fluctuations for an entire hour. Price volatility beyond fractional digits can cause even greater losses if there are interruptions to the price pipeline, slowing of liquidation activity or simultaneous deterioration of market depth.

Therefore, the fraction is essentially calibrated to deal with bad debts buffer rather than borrower protection levels. The proposal sets BTC-related thresholds below the model upper bound in order to cover risks that cannot be captured by historical data such as depth, cap and concentration; ETH has a smaller reserve space whereas WETH has already touched the top of the model.

For Aave's governance, the core issue was whether liquidation performance under existing parameters would be sufficient to support future positions operating in a location closer to the clearing line. To borrowers, this transaction looked promising with an increase of 8 percentage points and price buffers shrinking from about 6.4% to 4.7%. From a contractual standpoint, historical data showed that liquidation disposals were swift; two stress periods did not extend beyond reviewing collateral and arguments were sufficiently relative; but historical execution could not measure extreme events occurring at the same time as system outages.

According to the forum information, if Snapshot votes it will be submitted as an AIP for implementation. 81% of Bitcoin LTVs remain in the governance proposal phase until the vote is completed and officially landed.