The conventional approach to measuring lending protocol risk—Collateral at Risk (CaR)—answers a straightforward question: what value of positions would face liquidation under a given market scenario? But this metric tells only half the story. A new analytical framework called Liquidity-at-Risk (LaR) asks the more pressing follow-up: can decentralized exchange markets actually absorb these liquidations without slippage destroying the arbitrage economics that incentivize liquidators to execute?
This distinction matters enormously during systemic stress. A developer built a live dashboard tracking approximately 25,000 positions across Aave V3 and Compound V3, using real-time aggregator quotes from Uniswap, Curve, Balancer, and other venues to measure whether sufficient depth exists to absorb projected liquidations. The elegant threshold: when slippage on a liquidation trade exceeds the protocol's liquidation bonus, the transaction becomes economically irrational for any liquidator to execute. Using this framework, recent stress scenarios reveal startling gaps. If liquid staking token prices depreciated by 5 percent against ETH, weETH positions would face $1.7 billion in liquidations against only $19 million of available DEX liquidity—a shortfall of $1.68 billion. Similar dynamics plague wstETH, osETH, and rsETH, with aggregate LST shortfalls reaching $3.43 billion.
E-Mode markets deserve particular attention here. These high-leverage positions remain theoretically insulated from correlated crashes, since both collateral and borrowed assets move together, leaving health factors unchanged. But they are acutely vulnerable to depegging events where one side of the pair loses parity while the other holds firm. Most LST E-Mode positions offer only 1 percent liquidation bonuses, leaving minimal room for slippage before liquidators abandon unprofitable executions. A separate Ethena (ENA) depeg scenario illustrates this: $1 billion in collateral at risk, with a $939 million liquidity shortfall—a near-total failure of market absorption capacity.
The dashboard updates every 15 minutes and documents its methodology transparently, though users should note inherent limitations: the model captures only first-order liquidations without cascade effects, DEX snapshots age continuously, and subgraph price feeds may lag by several blocks. These caveats do not invalidate the framework—they simply anchor analysis in realistic constraints. As Aave and competing protocols scale, understanding whether liquidation mechanisms can function at all becomes as important as calculating when they would theoretically trigger.