Replication package for Systemic Liquidity Breakdown: Failure Boundaries, Shock Magnitude, and Cross-Sectional Geometry

This software and replication archive accompanies the study “Systemic Liquidity Breakdown: Failure Boundaries, Shock Magnitude, and Cross-Sectional Geometry.” It contains output-free computational code, frozen design objects, non-proprietary aggregate results, integrity hashes, and reproducibility documentation supporting the paper’s calibrated systemic liquidity boundary, matched state–shock transport, magnitude–geometry decomposition, robustness analyses, and external portability exercises. The final external conditional-reference analysis evaluates a non-overlapping 369-security universe. Each of 18 systemic events is compared with 20 unique same-year controls selected only from prior Bloomberg state variables \((C,S,\mathrm{SLM})\), with no control reuse. The mean conditional upper-tail shock percentile is 0.6444; 14 of 18 events exceed the matched-control median; and the exact one-sided matched-set reference probability is \(p=0.0222\). Licensed Bloomberg source observations and provider-controlled raw market data are not redistributed. The archive supports code inspection, verification of frozen design and aggregate outputs, and full rerunning by researchers who independently possess or reconstruct the required source data. The external evidence concerns upper-tail liquidity-shock portability rather than full replication of the proprietary direct-quote mechanism.