Deleveraging Cascade

A blockchain-native deleveraging cascade centered on an algorithmic stablecoin.

• Large withdrawals from the Anchor Protocol began destabilizing UST's peg.
• The arbitrage mechanism between UST and LUNA triggered hyperinflation of LUNA supply.
• As LUNA's price collapsed, it destroyed the collateral backing for UST, accelerating the death spiral.
• The event wiped out ~$40B in market value and triggered contagion across the DeFi ecosystem.

A cascade driven by over-leveraged hedge funds and CeFi lenders.

• Three Arrows Capital (3AC) faced margin calls on massive leveraged positions (e.g., stETH).
• Their default triggered liquidations for lenders like Voyager Digital and Celsius.
• Celsius's insolvency and freeze on withdrawals created a panic, forcing asset sales.
• This propagated losses to other platforms and contributed to the crypto bear market of 2022.

A DeFi-specific deleveraging event caused by market volatility and network congestion.

• A 50% drop in ETH price triggered undercollateralized Vaults (CDPs).
• Network congestion on Ethereum delayed keeper bots from executing liquidations.
• Some liquidations were processed at the zero bid of 0 DAI, causing a system deficit of ~$4 million.
• This exposed flaws in the liquidation mechanism and oracle reliance during extreme volatility.

A cascade stemming from a centralized exchange's insolvency.

• Revelations about FTX's misuse of customer funds and balance sheet insolvency caused a bank run.
• FTX's collapse revealed deep interconnections, exposing counterparty risk for many trading firms and DeFi protocols (e.g., Solana DeFi).
• The resulting credit contraction and loss of trust forced widespread de-risking and asset sell-offs across the industry.

Common threads across all major deleveraging cascades:

• Excessive Leverage: The primary fuel, often hidden in complex financial products or promises of high yield.
• Asset-Liability Mismatch: Illiquid assets backing short-term liabilities (e.g., bank deposits, stablecoin redemptions).
• Price-Oracle Dependency: Liquidations and solvency are based on potentially manipulable or lagging price data.
• Contagion via Interconnectedness: Failure in one node (bank, fund, protocol) rapidly spreads due to shared exposures and counterparty risk.

Protocols impose a liquidation penalty (e.g., 5-15%) on the liquidated position, which serves two purposes: it compensates the liquidator for their work and acts as a deterrent against excessive leverage. The penalty is typically split between the liquidator and the protocol treasury, creating a market-driven incentive system to keep the system healthy. This mechanism helps ensure there is always economic motivation to liquidate underwater positions before they threaten protocol solvency.

To prevent a single large liquidation from crashing the oracle price, protocols implement gradual price impact models or circuit breakers. Instead of dumping the entire collateral at the current market price, liquidations are processed in smaller batches over time or at a discount to a time-weighted average price (TWAP). This slows the feedback loop, giving the market time to absorb the selling pressure and reducing the likelihood of triggering cascading liquidations in other positions.

A key architectural choice is between isolated margin and cross-margin (or shared collateral) systems. In isolated mode, the risk from a position is contained; its liquidation cannot directly drain collateral from a user's other positions. This limits contagion. Cross-margin systems are more capital efficient but riskier, as the failure of one position can jeopardize a user's entire portfolio. Protocols like Aave V3 allow users to toggle asset isolation for high-risk assets.

The Health Factor is a core risk metric (e.g., Total Collateral / Total Borrowed). Protocols set a liquidation threshold (e.g., 1.1) well above the insolvency point of 1.0. This creates a safety buffer, ensuring liquidations are triggered while the position still has sufficient collateral to cover the debt after penalties. This buffer gives liquidators time to act and prevents the protocol from accruing bad debt, which is a primary trigger for cascades.

When a cascade leads to bad debt (undercollateralized loans that couldn't be liquidated), some protocols have automated recapitalization mechanisms. For example, MakerDAO uses debt auctions (flip and flap auctions) to mint and sell new governance tokens (MKR) to cover the system's shortfall. This transfers the cost to token holders but ensures the protocol remains solvent and user deposits are made whole, breaking the cascade of insolvency.

Protocols can employ dynamic risk parameters that automatically adjust based on market volatility. For instance, Loan-to-Value (LTV) ratios and liquidation thresholds can be temporarily lowered during periods of high volatility, forcing users to maintain higher collateralization. This pre-emptive de-risking makes the system more resilient to sudden price drops. Governance or automated keepers monitor oracle feeds and volatility indices to trigger these adjustments.

The forced closure of an undercollateralized position, where collateral is automatically sold to repay a loan. This is the primary trigger for a deleveraging cascade.

• Mechanism: When a borrower's collateralization ratio falls below the liquidation threshold, a liquidator can repay part of the debt in exchange for the collateral at a discount.
• Cascade Link: Mass liquidations in volatile markets create sell pressure, driving asset prices down and triggering further liquidations.

A numerical representation of the safety of a borrowed position, calculated as (Collateral Value / Borrowed Value).

• Function: It determines how close a position is to being liquidated. A health factor dropping below 1.0 (or a protocol-specific threshold) makes the position eligible for liquidation.
• Market Sensitivity: This ratio is highly sensitive to the market price of the collateral asset. A sharp price drop causes the ratio to plummet across many positions simultaneously.

A self-reinforcing cycle where one liquidation causes asset prices to fall, triggering more liquidations. This is the core sequence of a broader deleveraging event.

• Positive Feedback Loop: Liquidations → Increased sell pressure → Lower collateral prices → More undercollateralized positions → More liquidations.
• Example: The 2020 "Black Thursday" event on MakerDAO, where a rapid ETH price drop triggered a cascade of vault liquidations.

In traditional finance, a broker's demand that an investor deposit more money or securities to cover a potential loss on a margin trade. This is the centralized equivalent of an on-chain liquidation event.

• Key Difference: In DeFi, liquidations are permissionless, automated, and executed by bots, making the process faster and less discretionary than a traditional margin call.
• Systemic Risk: Both concepts highlight the inherent risk of leveraged positions when asset values decline.

A theory where market participants' perceptions influence economic fundamentals, which in turn change perceptions. In crypto, this describes the feedback loop between price action and on-chain activity.

• DeFi Application: Falling prices (perception) trigger liquidations (fundamental on-chain action), which cause further price declines (new perception). This reflexivity accelerates a deleveraging cascade.
• Narrative Impact: Fear of a cascade can lead to pre-emptive deleveraging, becoming a self-fulfilling prophecy.

A critical distinction during a crisis. Insolvency means liabilities exceed assets. Illiquidity means assets cannot be sold quickly enough to meet obligations without significant loss.

• Cascade Dynamics: Initial liquidations are often driven by illiquidity (inability to post more collateral). If prices fall far enough, systemic insolvency can occur, where the total debt in the system exceeds the value of all collateral.
• Protocol Risk: Designing systems to withstand illiquidity shocks is key to preventing irreversible insolvency.