{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp11_mechanism_context_external20_20260804","research_id":"eoa_inverse_innovation_exp11_external_scrutiny_20260804","cell_id":"invariant_mode_decomposition_design__economics_finance","opaque_id":"invariant_mode_decomposition_design__economics_finance__B","search_lanes":{"direct_problem":{"queries":["dealer balance sheet common deleveraging asset sales funding liquidity systemic risk evidence","dealer balance sheet capacity market liquidity selloff VaR regulatory capital ratios","site:bis.org liquidity stress testing system-wide simultaneous asset sales funding withdrawal banks supervisory principles"],"source_ids":["S2","S3","S4","S7"],"no_result_note":null},"closest_prior_art":{"queries":["dynamic mode decomposition financial systemic risk eigenvectors balance sheets liquidity","Koopman dynamic mode decomposition bank balance sheets systemic risk early warning out of sample","spectral modes financial stress early warning liquidity eigenvalues dealer","system-wide liquidity stress testing common asset sales dealer balance sheets"],"source_ids":["S1","S4","S7"],"no_result_note":null},"historical_terminology":{"queries":["historical terminology concerted deleveraging market liquidity spiral dealers","Brunnermeier Pedersen market liquidity and funding liquidity 2009 PDF","Adrian Shin liquidity and leverage procyclical dealer balance sheets 2010 PDF"],"source_ids":["S2","S3"],"no_result_note":null},"products_practices_standards":{"queries":["site:federalreserve.gov supervisory liquidity stress test dealers FR 2052a liquidity monitoring systemic","site:federalreserve.gov FR 2052a complex institution liquidity monitoring report dealer data","site:federalreserve.gov authority require liquidity buffer bank holding company liquidity stress testing","system-wide liquidity stress testing common asset sales dealer balance sheets product platform"],"source_ids":["S4","S5","S6","S7"],"no_result_note":null},"non_english_regional":{"queries":["test de résistance systémique liquidité ventes d'actifs simultanées banques dealers financement Banque de France","systemweiter Liquiditätsstresstest gleichzeitige Vermögensverkäufe Banken Finanzierung","prueba de estrés de liquidez sistémica ventas simultáneas activos bancos financiación","金融 システム 流動性 ストレステスト 同時 資産売却 資金調達 銀行"],"source_ids":["S7"],"no_result_note":null},"composition_subproblems":{"queries":["eigenvector autoregression financial systemic risk modes liquidity stress prediction","stepped wedge randomized trial bank regulation liquidity buffer supervisory intervention","randomized rollout financial regulation banks causal liquidity requirement experiment","financial regulation randomized trial liquidity buffer banks HQLA funding maturity intervention"],"source_ids":["S1","S4","S6","S8"],"no_result_note":"No retained source documented the complete composition of a frozen dealer-panel transition operator, sealed modal alert comparison, and live cluster-randomized HQLA and maturity interventions."}},"sources":[{"source_id":"S1","title":"Evaluation of systemic risk in a financial system using dynamic mode decomposition","url":"https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6031928","publisher":"SSRN / Nara Institute of Science and Technology","date_or_year":"2026-01-07","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["Applies dynamic mode decomposition to banking statements and market data for major US financial institutions.","Approximates financial-system dynamics with a Koopman operator and defines instability using its eigenvalues.","Creates a close collision with the proposal's spectral detection mechanism, but does not report dealer-liquidity controls, randomized interventions, or the proposal's sealed comparative evaluation."]},{"source_id":"S2","title":"Market Liquidity and Funding Liquidity","url":"https://eprints.lse.ac.uk/24478/1/dp580.pdf","publisher":"London School of Economics","date_or_year":"2007 working-paper version; published 2009","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["Links dealers' and traders' funding constraints to market liquidity.","Shows theoretically that margins, losses, forced position reductions, and market illiquidity can reinforce one another in liquidity spirals.","Explains common liquidity deterioration across securities, supplying older terminology for the proposed joint deleveraging problem."]},{"source_id":"S3","title":"Dealer Balance Sheet Capacity and Market Liquidity during the 2013 Selloff in Fixed-income Markets","url":"https://www.federalreserve.gov/econresdata/notes/feds-notes/2013/dealer-balance-sheet-capacity-and-market-liquidity-during-the-2013-selloff-in-fixed-income-markets-20131016.html","publisher":"Board of Governors of the Federal Reserve System","date_or_year":"2013-10-16","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["Documents a period in which market liquidity deteriorated while dealers reduced fixed-income positions.","Finds that dealers with more room relative to VaR and capital constraints sold more, inconsistent with a simple binding-ratio explanation.","Supports the possibility that firm-level regulatory ratios alone can miss coordinated balance-sheet retrenchment driven by risk appetite."]},{"source_id":"S4","title":"Liquidity stress tests for banks – range of practices and possible developments","url":"https://www.bis.org/fsi/publ/insights59.htm","publisher":"Bank for International Settlements, Financial Stability Institute","date_or_year":"2024-10-11","source_type":"OFFICIAL_GUIDANCE","language":"English","claims_supported":["Reviews established authority-run institution-level, sector-wide, and bank–NBFI liquidity stress-test practices.","Identifies propagation, management responses, interactions, second-round effects, contagion, data, and modelling as central issues.","Shows that the broad problem and system-wide stress-testing response are established, while leaving room for a distinct empirically estimated modal alert."]},{"source_id":"S5","title":"FR 2052a Complex Institution Liquidity Monitoring Report","url":"https://www.federalreserve.gov/apps/reportingforms/Report/Index/FR_2052a","publisher":"Board of Governors of the Federal Reserve System","date_or_year":"Current form, accessed 2026-08-04","source_type":"OFFICIAL_GUIDANCE","language":"English","claims_supported":["Collects detailed assets, liabilities, funding activities, contingent liabilities, maturities, and material-entity information.","Includes broker-dealer entities and supports both firm-specific surveillance and aggregate macroprudential supervision.","Identifies the Federal Reserve and covered large banking organizations as plausible authorizer and adopter, while not establishing that every proposed exposure, sale, margin, and market-outcome variable is complete and consistently measured."]},{"source_id":"S6","title":"12 CFR 252.35 — Liquidity Stress Testing and Buffer Requirements","url":"https://www.federalreserve.gov/frrs/regulations/section-25235-liquidity-stress-testing-and-buffer-requirements.htm","publisher":"Board of Governors of the Federal Reserve System","date_or_year":"Current regulation, accessed 2026-08-04","source_type":"OFFICIAL_STANDARD","language":"English","claims_supported":["Requires covered bank holding companies to conduct liquidity stress tests at least monthly or quarterly, depending on category.","Requires tests to incorporate adverse actions by other stressed market participants and combined market and idiosyncratic scenarios.","Requires a 30-day liquidity buffer composed of unencumbered highly liquid assets, establishing the HQLA lever as an existing regulatory practice.","Provides authority for stress testing and buffers but does not itself authorize a randomized supervisory trial or modal targeting."]},{"source_id":"S7","title":"La Banque de France, l'ACPR et l'AMF publient un rapport méthodologique sur leur premier test de résistance portant sur l'ensemble du système financier","url":"https://www.banque-france.fr/fr/communiques-de-presse/la-banque-de-france-lacpr-et-lamf-publient-un-rapport-methodologique-sur-leur-premier-test-de","publisher":"Banque de France / ACPR / AMF","date_or_year":"2026-06-17","source_type":"OFFICIAL_GUIDANCE","language":"French","claims_supported":["Describes a voluntary French system-wide pilot covering banks, insurers, asset managers, and a clearing house.","Explicitly examines simultaneous sales of similar assets, common demands on liquidity sources, margin calls, repos, counterparty interactions, and second-round feedback.","Demonstrates that identifying collectively inconsistent defensive responses is already an active supervisory practice, though without the proposal's learned eigenmodes or randomized modal interventions.","Identifies concrete regional supervisory adopters and authorizers."]},{"source_id":"S8","title":"The costs and benefits of liquidity regulations: Lessons from an idle monetary policy tool","url":"https://www.federalreserve.gov/econres/feds/the-costs-and-benefits-of-liquidity-regulations-lessons-from-an-idle-monetary-policy-tool.htm","publisher":"Board of Governors of the Federal Reserve System","date_or_year":"2019-05","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["Uses a regression-kink design around an exogenous threshold to estimate effects of mandated HQLA increases.","Finds that higher HQLA requirements reduced bank failure odds but also reduced credit supply and profitability.","Supports both the plausibility of the buffer lever and the proposal's concern about costs shifted to clients or other activities."]}],"problem_evidence":{"status":"PARTLY_SUPPORTED","finding":"The underlying problem is well supported: funding constraints, margin and loss spirals, simultaneous asset sales, common liquidity-source demands, and dealer retrenchment can jointly impair market liquidity, and firm-level constraint measures need not explain which dealers sell most. The stronger formulation—a recurring, stable dealer-panel eigenmode that remains hidden to both firm thresholds and network stress tests—has not yet been demonstrated.","source_ids":["S2","S3","S4","S7"],"uncertainty":"The evidence includes theory, individual historical episodes, supervisory reviews, and an exploratory system-wide exercise; it does not establish persistence, eigenvector stability, or incremental sealed-period warning for the proposed state vector."},"adopter_evidence":{"status":"SUPPORTED","finding":"The Federal Reserve is an identifiable US authorizer and adopter: it collects FR 2052a liquidity data, supervises covered large organizations, and administers binding liquidity stress-test and buffer requirements. Banque de France, ACPR, and AMF provide independent regional examples of authorities conducting voluntary system-wide liquidity exercises.","source_ids":["S5","S6","S7"],"uncertainty":"These sources establish institutional authority over reporting, stress testing, and some liquidity requirements, but not authority for the proposed randomized dealer-level changes; trial authority, firm consent, and cross-agency jurisdiction would require separate legal review."},"implementation_evidence":{"status":"PARTLY_SUPPORTED","finding":"The main ingredients are feasible in isolation: supervisory liquidity data include entity, funding, maturity, asset, and contingent-liability detail; DMD has been applied to financial-institution statements and market data; authorities already run system-wide liquidity tests; and causal evidence exists for HQLA requirements. No source establishes the completeness and frequency of the exact fixed-dealer panel, reliable estimation of stable one-month modes, or feasibility of the proposed 2-by-2 stepped-wedge live trial.","source_ids":["S1","S4","S5","S6","S8"],"uncertainty":"Confidential-data access, changing dealer panels and legal entities, strategic anticipation, interference, rare stress periods, nonlinearity, and legal limits on randomized supervisory treatment remain material implementation risks."},"prior_art":{"disposition":"ADJACENT_PRIOR_ART","closest_analogues":[{"name":"Dynamic-mode/Koopman systemic-risk metric for major US financial institutions","source_ids":["S1"],"same_problem":true,"same_causal_lever":false,"overlap":"Uses financial-institution statements plus market data, estimates dynamic modes, and classifies instability through Koopman-operator eigenvalues.","remaining_difference":"Does not target dealer funding-withdrawal and asset-sale states, report the proposed frozen train/validation/test gates, compare against both firm thresholds and a network stress test under a common alert budget, or randomize HQLA and maturity controls."},{"name":"French system-wide liquidity stress-test pilot","source_ids":["S7"],"same_problem":true,"same_causal_lever":false,"overlap":"Detects simultaneous asset-sale plans, common calls on liquidity, counterparty inconsistencies, management reactions, and second-round systemic feedback.","remaining_difference":"Uses scenario responses and aggregation rather than an empirically learned invariant-mode basis, and imposes no live modal control."},{"name":"Established sector-wide liquidity stress-test practice","source_ids":["S4","S6"],"same_problem":true,"same_causal_lever":false,"overlap":"Covers market-wide shocks, management actions, propagation, interactions, contagion, and stressed funding needs.","remaining_difference":"Does not define risk as an out-of-sample stable eigenmode or test whether a modal alert adds warning beyond matched comparators."},{"name":"Regulatory liquidity buffers and causal evidence on HQLA requirements","source_ids":["S6","S8"],"same_problem":true,"same_causal_lever":true,"overlap":"HQLA buffers are already required, and quasi-experimental evidence evaluates their prudential benefits and economic costs.","remaining_difference":"The buffer is not assigned according to a learned mode, the evidence is not a consenting-dealer randomized trial, and it does not test effects on a preregistered mode amplitude or gain."},{"name":"Funding–market-liquidity spiral theory and dealer-selloff evidence","source_ids":["S2","S3"],"same_problem":true,"same_causal_lever":false,"overlap":"Provides the established mechanism by which funding conditions, margins, risk appetite, asset sales, and common market illiquidity reinforce each other.","remaining_difference":"Does not estimate or prospectively validate a fixed joint-state transition operator or test modal interventions."}],"contrastive_claim_remaining":"In a fixed dealer panel, a fully frozen one-month local operator will reveal a stable joint funding-withdrawal/asset-sale mode that improves final-block precision-recall area by at least 0.05 and median warning lead time by at least five business days versus both matched firm-level thresholds and a frozen network stress test; conditional on separate approval, randomized HQLA and maturity extensions will then reduce that mode's amplitude or next-month gain by at least 10% on the stated timetable without placebo effects or subgroup harm.","contrastive_claim_falsifier":"The claim is falsified if the mode fails any frozen reconstruction, bootstrap-direction, spectral-gap, predictive-improvement, or lead-time gate; or, if a separately authorized trial occurs, if neither intervention achieves its preregistered timing and 10% effect with a confidence interval excluding zero, placebo tests respond, assignment is contaminated, liquidity worsens, or the client-cost harm threshold is crossed.","confidence":"HIGH","search_limitations":"The bounded search covered six adversarial lanes and eight direct sources across SSRN, LSE, the Federal Reserve, BIS, and French authorities. The SSRN full-text server returned HTTP 403, so S1 was assessed from its publicly indexed abstract and bibliographic page. No proprietary supervisory datasets, patent databases, paid literature databases, internal model documentation, or legal opinions were searched. The search cannot establish world novelty, patentability, freedom to operate, market size, or realized impact."},"researchability_gates":{"externally_supported_problem":{"status":"PASS","rationale":"Independent theory, dealer-level evidence, supervisory reviews, and a regional system-wide pilot support the existence of coupled funding, asset-sale, and market-liquidity feedback even though the specific stable eigenmode remains unproven.","source_ids":["S2","S3","S4","S7"]},"identifiable_adopter_or_authorizer":{"status":"PASS","rationale":"The Federal Reserve is identifiable through its reporting, surveillance, stress-testing, and buffer roles; Banque de France, ACPR, and AMF independently demonstrate comparable system-wide supervisory adoption.","source_ids":["S5","S6","S7"]},"distinct_testable_incremental_claim":{"status":"PASS","rationale":"Close spectral and supervisory prior art exists, but no retained source combines a dealer-specific frozen modal alert, matched out-of-sample comparisons, and randomized mode-targeted HQLA and maturity interventions. The remaining performance and treatment claims have explicit quantitative falsifiers.","source_ids":["S1","S4","S6","S7","S8"]},"bounded_next_evidence_step":{"status":"PASS","rationale":"A legal/data audit followed by the frozen retrospective shadow evaluation is bounded, non-live, reversible, and capable of rejecting the key predictive claim before any intervention.","source_ids":["S4","S5","S6"]},"no_unresolved_safety_or_authority_stop":{"status":"PASS","rationale":"No authority or safety issue stops the retrospective shadow evaluation when confined to authorized analysts and confidential data. Live randomization remains expressly contingent on separate legal authority, consent where required, independent review, contamination checks, and harm stopping rules; it is not part of the authorized next step.","source_ids":["S5","S6","S7","S8"]},"adequate_search_evidence":{"status":"PASS","rationale":"The search used all six required lanes, exactly eight direct sources, five independent institutional publishing channels, multiple primary research sources, and multiple official or standards sources, with adversarial attention to exact DMD prior art, established stress-testing practice, non-English regional terminology, data, authority, and intervention side effects.","source_ids":["S1","S2","S3","S4","S5","S6","S7","S8"]}},"strict_success":true,"screen_survival":true,"remaining_research_value":"MODERATE","recommended_next_step":"Do not initiate live controls. First obtain a written legal and data-access determination, audit FR 2052a and linked market data for a stable dealer/entity panel and the preregistered variables, then run the frozen retrospective train/validation/sealed-test shadow evaluation against both comparators. Stop if any fidelity, stability, gap, predictive, confidentiality, or subgroup-relevance gate fails; only if all pass should an independent body assess whether a separately authorized and consenting pilot is lawful and feasible.","world_novelty_boundary":"This result establishes only that the bounded public search found adjacent prior art while leaving a distinct falsifiable empirical composition. It does not establish world novelty, patentability, freedom to operate, market size, legal authority for live experimentation, or realized financial-stability impact."}