Cryptocurrency systems are prone to fast, non-linear contagion because on-chain linkages and off-chain leverage interact across jurisdictions and platforms. Studies show the most revealing stress tests combine network balance-sheet models, fire-sale liquidity simulations, and agent-based behavior to expose realistic contagion pathways rather than relying on simple correlation shocks. Evidence from the systemic-risk literature supports this integrated approach.
Network and balance-sheet methods
Network-based approaches map exposures between entities and propagate defaults through obligations. DebtRank developed by Stefano Battiston University of Zurich quantifies how distress travels via network centrality and second-round effects, making it well suited to crypto ecosystems where custody and derivative exposure concentrate risk. Pure clearing models that compute direct loss propagation are valuable but can understate amplification unless they incorporate cross-holdings, margining rules, and rehypothecation chains typical in decentralized finance.
Liquidity and agent-based simulations
Liquidity stress tests capture the fire-sale channel that often triggers wider collapse. Work on funding liquidity and market liquidity interactions by Markus Brunnermeier Princeton University highlights how margin calls and asset price drops create feedback loops. In crypto, thin order books, concentrated market-making, and on-chain settlement delays amplify these spirals. Agent-based models add behavioral realism by simulating heterogeneous actors, runs on exchanges, and rational panic that generate tail outcomes unseen in Gaussian shock frameworks.
Relevance arises from causes rooted in crypto market structure. Concentrated staking providers, opaque counterparty exposures, leverage in perpetual futures, and geographic clustering of mining and custodial infrastructure create territorial and cultural vectors for contagion. Consequences extend beyond price loss: network outages, cross-border regulatory frictions, and environmental impacts from abrupt mining shutdowns can cascade to traditional finance where intermediation exists.
Practical stress testing therefore layers methods and data. On-chain transaction graphs and smart-contract dependency maps feed network contagion models while order-book and funding-rate histories drive liquidity simulations. Validated scenarios should include localized infrastructure shocks, governance attacks, and coordinated deleveraging episodes. No single method fully captures systemic crypto contagion, but integrated, validated frameworks that combine network topology, behavioral dynamics, and liquidity channels best reveal realistic pathways and inform targeted mitigations.