When Cross-Venue Agreement Is Not Price Discovery: Disclosure Frontiers for 24/7 Equity-Perpetual Oracles

Battle Count: 80

Highly relevant for traders and risk managers in crypto perpetual markets. The paper directly addresses how closed-window marks (which set margin, funding, and liquidation for hours at a time) are formed and whether cross-venue agreement reflects genuine price discovery or endogenous feedback. The impossibility results warn that standard price-discovery estimators (lead-lag, information shares, Granger causality) have power equal to size in this setting. The disclosure frontier provides actionable guidance for evaluating venue oracle quality and managing tail risk from mis-anchored marks during cash-market closures.

The theoretical framework (fixed-point equations, identification theorems, path-law reparameterization) requires graduate-level econometrics and linear algebra. The empirical implementation involves estimating reduced forms, constructing admissibility sets, running bootstrap tests, and computing projection distances. The simulation code is provided with fixed seeds. However, the identification theory is abstract and applying it to new venues requires careful reading of rule disclosures and handling of provenance-ambiguous inputs.

The paper provides frozen public data snapshots with SHA-256 hashes, fixed-seed simulation scripts (figures/make_simulation.py, figures/make_experiments.py), and reproducible component snapshots under artifact/source_snapshots/. However, venue rules are date-sensitive help-center disclosures that may change, and some provenance-ambiguous inputs require dual coding. The empirical samples are small (5 reopen days, 8-9 multi-day closures).