PQOD · Asile — WP-01 — Closed-Loop Steering for European Asylum and Return Policy: A Preregistered Quantum-Operator Audit of the EU Response Chain (2013–2026)

White paper of the PQOD programme (series « PQOD · Asile », WP-01), September 2026 — EMBARGOED until 2027-09-16. We audit the European asylum and return policy as a dynamical system: the response chain (applications → first-instance decisions → protection → return decisions → executed returns → resettlements) is modelled by transition operators estimated on real administrative data (Eurostat asylum and return series, 158 series / 5 685 points; UNHCR; Frontex; INSEE; OFPRA), under a preregistered and hashed protocol (SHA-256, seed 20260915, 10 000 bootstrap replicates) with a prior-art Blue Ocean gate (5 GO). Findings: the EU return execution rate spent five years below the 30 % lock threshold (2020–2024, trough 20.8 % in 2022, 18 of 21 quarters locked); the 2025 exit (32.1 %) is too recent to be durable — P(lock in 2030) = 0.360 [0.123 ; 0.683]; a 50 % target has only a 9.3 % chance under free dynamics; the structural shortfall is 17.9 points ≈ 87 170 returns per year; the operational rectification dose is 3.59 points per year ≈ 17 470 additional executed returns per year (robust forcing f = 0.320 ≈ 9 250; target forcing f = 0.607 ≈ 17 890 — a regime choice, not a marginal adjustment). Regime breaks are dated by CUSUM after Holm correction (argmax 2021-12, installed break; COVID is a level shock without regime change, D = 0.013). Instrument order is refuted at the frozen criterion (5/5 distances exclude zero but mean sequence effect q = +0.061, p = 0.035) and reported with the same precision; the spectral delay predictor is refuted for the third consecutive campaign and permanently abandoned. All figures reproducible: preregistered thresholds, sourced public data, executable code.