Abstract
Auto-deleveraging (ADL) mechanisms are a critical yet understudied component of risk management on cryptocurrency futures exchanges. When available margin and other loss-absorbing resources are insufficient to cover losses following large price moves, exchanges reduce positions and socialize losses among solvent participants via rule-based ADL protocols. We formulate ADL as an optimization problem that minimizes the exchange's risk of loss arising from future equity shortfalls. In a single-asset, isolated-margin setting, we show that under a risk-neutral expected loss objective the unique optimal policy minimizes the maximum leverage among participants. The resulting design has a transparent structure: positions are reduced first for the most highly levered accounts, and leverage is progressively equalized via a water-filling (or "leverage-draining") rule. This policy is distribution-free, wash-trade resistant, Sybil resistant, and path-independent. It provides a canonical and implementable benchmark for ADL design and clarifies the economic logic underlying queue-based mechanisms used in practice. We further study the multi-asset, cross-margin setting, where the ADL problem becomes genuinely multi-dimensional: the exchange must allocate a vector of required reductions across accounts with portfolios exposed to correlated price moves. We show that under an expected-loss objective the problem remains separable across accounts after introducing asset-level shadow prices, yielding a scalable numerical method. We observe that naive gross leverage can be misleading in this context as it ignores hedging within portfolios. When asset prices are driven by a single dominant risk factor, the optimal policy again takes a water-filling form, but now in a factor-adjusted notion of leverage, so that more effectively hedged portfolios are deleveraged less aggressively.
Introduction
In the high-volatility environment of cryptocurrency futures, extreme price movements can lead to liquidations that exhaust a trader's margin. Typically, an exchange's insurance fund absorbs these losses to maintain market integrity. However, during "black swan" events, these funds can be depleted. When this occurs, exchanges resort to Auto-Deleveraging (ADL)—a "nuclear option" where the positions of solvent, profitable traders are forcibly closed to offset the losses of bankrupt ones.
Despite its importance, current ADL mechanisms used by major exchanges are largely heuristic. Most employ a priority queue that ranks accounts based on a combination of profit and leverage. These methods lack a rigorous economic foundation and are often vulnerable to manipulation, such as Sybil attacks, where a trader splits a large position into multiple smaller accounts to avoid being at the front of the ADL queue. This paper moves ADL from these ad-hoc heuristics to a rigorous optimization framework designed to minimize the exchange's future risk.
The Single-Asset Setting: Minimax Leverage
The research formalizes ADL as a risk-minimization problem. In a single-asset, isolated-margin environment, the exchange seeks an allocation of position reductions that minimizes the risk of future equity shortfalls. The authors prove that under a risk-neutral "expected loss" objective, the unique optimal policy is minimax leverage.
This policy is implemented via a "water-filling" (or leverage-draining) algorithm. The exchange identifies the most highly levered account and reduces its position until its leverage matches that of the next most levered account. The process continues, reducing both accounts together until they match the third, and so on, until the total required reduction is achieved. This approach is not only mathematically optimal but also provides several critical properties:
- Sybil Resistance: Because the rule targets maximum leverage, splitting a position across multiple accounts does not change the aggregate leverage, offering no protection to the trader.
- Wash-Trade Resistance: The policy cannot be gamed by artificial trading activity intended to alter queue priority.
- Path Independence: Applying ADL in multiple small steps yields the same result as a single large step, ensuring consistency during volatile periods.
Multi-Asset and Cross-Margin Complexity
When traders hold portfolios of multiple correlated assets (e.g., BTC and ETH) under a cross-margin system, the ADL problem becomes multi-dimensional. The exchange must decide how to allocate reductions across a vector of assets. A significant challenge here is that "gross leverage" (total position size divided by equity) becomes a misleading metric because it ignores the risk-reducing effects of hedging.
To solve this, the authors utilize convex duality to introduce "shadow prices" for each asset. This mathematical approach allows a massive, global optimization problem to be decomposed into independent sub-problems for each account. This decomposition is vital for scalability, allowing exchanges to process ADL for millions of users in parallel.
Factor-Adjusted Leverage
In scenarios where asset prices are driven by a single dominant risk factor (such as general crypto market sentiment), the authors derive an elegant extension of the water-filling rule. In this context, the rule is applied to factor-adjusted leverage. This metric correctly accounts for correlations, ensuring that well-hedged portfolios are deleveraged less aggressively than those with high directional exposure to the dominant risk factor.
Conclusion
By shifting the focus from heuristic queues to risk-based optimization, this research provides a new blueprint for market stability. The minimax leverage policy and the resulting water-filling algorithm offer a transparent, robust, and computationally efficient benchmark for exchange design. While the model currently assumes fixed execution prices, it clarifies the economic logic necessary to protect exchanges and their participants during periods of extreme market stress.
Summary updated