{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp04_retrieval_first_paired20_20260802","cell_id":"invariant_mode_decomposition_design__economics_finance","round_index":0,"assessments":[{"hypothesis_id":"H1","search_queries":["bank intraday liquidity eigenvalue eigenvector stress testing liquidity allocation","site:bis.org intraday liquidity monitoring tools BCBS 248","spectral early warning bank liquidity instability eigenvector"],"sources":[{"source_id":"H1-S1","title":"Monitoring tools for intraday liquidity management","publisher":"Basel Committee on Banking Supervision","url":"https://www.bis.org/publ/bcbs248.htm","source_class":"STANDARD","claims_supported":["BCBS specifies seven quantitative intraday-liquidity monitoring tools, stress scenarios, application issues, and reporting requirements.","The tools are for monitoring and do not prescribe eigenmode-based allocation of contingent liquidity."]},{"source_id":"H1-S2","title":"Sound practices for intraday liquidity risk management","publisher":"European Central Bank Banking Supervision","url":"https://www.bankingsupervision.europa.eu/press/supervisory-newsletters/newsletter/2024/html/ssm.nl241113_2.en.html","source_class":"OFFICIAL_GUIDANCE","claims_supported":["Supervised banks use real-time monitoring, automated alerts, active outflow management, and stress testing to quantify intraday buffers.","ECB identifies interacting drivers such as payment behavior, collateral, margins, currencies, entities, and time-specific obligations, but does not describe modal instability-based buffer allocation."]},{"source_id":"H1-S3","title":"Early warning of systemic risk in global banking: eigen-pair R number for financial contagion and market price-based methods","publisher":"University of Essex Research Repository / Annals of Operations Research","url":"https://repository.essex.ac.uk/30689/","source_class":"PRIMARY_RESEARCH","claims_supported":["A spectral eigen-pair method treats banking-network failure as a dynamical-system stability problem and supplies an early-warning tipping-point measure.","The application is system-level contagion monitoring rather than bank-hour intraday treasury allocation."]}],"closest_analogue":"Eigen-pair systemic-risk early warning combined with ECB-style real-time intraday liquidity stress monitoring and buffer management.","overlap":"Existing work already combines granular liquidity drivers, stress scenarios, dynamic feedback, automated alerts, and spectral instability measures in banking-risk settings.","remaining_difference":"The testable difference is an operational bank-hour policy that estimates a local withdrawal-collateral-payment mode and allocates contingent liquidity to its exposure loadings. A retrospective comparison can test the claimed one-business-day lead and false-alarm bound against standard liquidity thresholds.","classification":"POSSIBLE_DISTINCTION","disposition":"ADVANCE","rationale":"The ingredients are established, but the shallow search did not find a direct analogue using locally unstable modes to allocate a single bank's intraday liquidity buffer. The operational allocation rule and stated benchmark remain concrete and testable."},{"hypothesis_id":"H2","search_queries":["debt restructuring creditor network centrality holdout sequencing consent solicitation","sovereign debt restructuring creditor coordination network centrality holdouts collective action clauses","corporate restructuring stakeholder network centrality creditor negotiation sequencing"],"sources":[{"source_id":"H2-S1","title":"WDR 2022 Chapter 5: Managing sovereign debt","publisher":"World Bank","url":"https://www.worldbank.org/en/publication/wdr2022/brief/chapter-5-managing-sovereign-debt","source_class":"OFFICIAL_GUIDANCE","claims_supported":["Creditor coordination and holdout risk materially delay restructurings.","Transparency, preemptive negotiation, and collective action clauses are established tools for accelerating resolution."]},{"source_id":"H2-S2","title":"Sovereign Debt Information","publisher":"International Capital Market Association","url":"https://www.icmagroup.org/resources-2/Sovereign-Debt-Information/?showiframe=true","source_class":"STANDARD","claims_supported":["ICMA publishes model collective-action, pari-passu, creditor-engagement, and majority-voting provisions intended to facilitate restructurings.","Existing contractual mechanisms organize creditor consent and constrain minority holdouts without using spectral centrality."]},{"source_id":"H2-S3","title":"The contractual approach to sovereign debt restructuring","publisher":"Bank of England","url":"https://www.bankofengland.co.uk/working-paper/2011/the-contractural-approach-to-sovereign-debt-restructuring","source_class":"PRIMARY_RESEARCH","claims_supported":["Collective-action clauses and creditor heterogeneity are formally analyzed as mechanisms for reducing litigation and restructuring inefficiency.","The analogue focuses on contractual voting and claim heterogeneity rather than network-ranked outreach sequencing."]}],"closest_analogue":"Creditor-committee sequencing and collective-action-clause design that identifies blocking positions and coordinates representative creditors.","overlap":"The established workflow maps claims and voting rights, identifies holdout or blocking power, organizes representative creditor groups, and sequences engagement to obtain consent.","remaining_difference":"The bounded research question is whether any restructuring adviser or empirical study has represented interdependent consent rights as a graph and used eigenvector or related spectral sensitivity—rather than claim size, seniority, class, or blocking percentage—to select the next creditor for outreach, side payment, or rights purchase.","classification":"POSSIBLE_DISTINCTION","disposition":"ADVANCE","rationale":"Holdout identification and negotiation sequencing are mature, but the retrieved direct sources did not show spectral centrality being used as the operational sequencing rule. The proposed comparison against claim-size ranking is bounded and falsifiable."},{"hypothesis_id":"H3","search_queries":["thin file SME lending PCA feature reduction alternative data underwriting","principal component analysis credit scoring feature selection loan applications paper","human in the loop credit underwriting residual review machine learning"],"sources":[{"source_id":"H3-S1","title":"A knowledge-informed neural network integrating fuzzy AHP and PCA for SME credit risk assessment","publisher":"Scientific Reports","url":"https://www.nature.com/articles/s41598-025-21441-4","source_class":"PRIMARY_RESEARCH","claims_supported":["The study applies PCA dimensionality reduction to information-scarce SME credit assessment and combines it with expert-derived risk judgments.","It presents the resulting model as an operational decision-support tool."]},{"source_id":"H3-S2","title":"Principal Component Analysis and Factor Analysis for Feature Selection in Credit Rating","publisher":"arXiv","url":"https://arxiv.org/abs/2011.09137","source_class":"PRIMARY_RESEARCH","claims_supported":["PCA and factor analysis are used to summarize correlated credit variables and obtain smaller feature sets.","The study explicitly compares reduced sets for predictive accuracy and reports that fewer factors can retain nearly the same accuracy."]},{"source_id":"H3-S3","title":"Consumer Financial Protection Circular 2022-03: Adverse action notification requirements in connection with credit decisions based on complex algorithms","publisher":"Consumer Financial Protection Bureau","url":"https://www.consumerfinance.gov/compliance/circulars/circular-2022-03-adverse-action-notification-requirements-in-connection-with-credit-decisions-based-on-complex-algorithms/","source_class":"GOVERNMENT_OR_REGULATOR","claims_supported":["Creditors using complex algorithms must still provide specific and accurate principal reasons for adverse actions.","Regulatory requirements already force residual exceptions, inadequate information, and model opacity to remain visible rather than being silently discarded."]}],"closest_analogue":"PCA-reduced SME credit assessment with expert or human decision support and regulatory exception handling.","overlap":"Dimensional reduction of correlated underwriting evidence, preservation of predictive performance, expert involvement, and escalation of unclear or inadequately explained decisions are all established components.","remaining_difference":"The narrower difference is measuring applicant burden directly and routing applications to manual review specifically when reconstruction residuals are structured or subgroup-specific, rather than when a score is merely borderline or unexplained.","classification":"OBVIOUS_COLLISION","disposition":"REJECT","rationale":"The core intervention is a direct assembly of established PCA credit-feature reduction, accuracy validation, expert review, and regulated explanation safeguards. Residual-triggered routing is a modest implementation rule, insufficient at this screening stage to separate the hypothesis from the crowded analogue."},{"hypothesis_id":"H4","search_queries":["policy design anticipatory behavioral response multiple margins bundle optimization Lucas critique","tax subsidy policy bundle endogenous behavioral responses sufficient statistics multi-dimensional adjustment","announcement effects anticipatory response dynamic factor model policy intervention"],"sources":[{"source_id":"H4-S1","title":"Anatomy of Welfare Reform Evaluation: Announcement and Implementation Effects","publisher":"University College London","url":"https://www.ucl.ac.uk/~uctp39a/BFK_IZAdp6050.pdf","source_class":"PRIMARY_RESEARCH","claims_supported":["Forward-looking agents can change behavior between policy announcement and implementation.","The empirical welfare-reform application finds substantial announcement effects and explains how they alter conventional impact estimates."]},{"source_id":"H4-S2","title":"Multi-Dimensional Pass-Through and Welfare Measures under Imperfect Competition","publisher":"arXiv","url":"https://arxiv.org/abs/1702.04967","source_class":"PRIMARY_RESEARCH","claims_supported":["The paper analyzes multiple policy instruments jointly and derives multidimensional pass-through and welfare measures.","It treats policy parameters as a vector and models cross-instrument effects rather than isolated elasticities."]},{"source_id":"H4-S3","title":"Estimation of Impulse-Response Functions with Dynamic Factor Models: A New Parametrization","publisher":"arXiv","url":"https://arxiv.org/abs/2202.00310","source_class":"PRIMARY_RESEARCH","claims_supported":["Dynamic factor models recover common response structure from high-dimensional macroeconomic data and estimate responses to monetary-policy shocks.","The paper emphasizes identification restrictions and observational equivalence, which constrain causal interpretation of recovered modes."]}],"closest_analogue":"Dynamic-factor policy-shock analysis combined with multidimensional pass-through optimization and announcement-effect estimation.","overlap":"Existing research already estimates common response factors, models anticipatory announcement effects, and evaluates interacting policy instruments through multidimensional pass-through rather than one-margin-at-a-time analysis.","remaining_difference":"The bounded unresolved question is whether any deployed policy-design study estimates announcement-window offset factors and then prospectively chooses a tax, subsidy, or credit bundle to minimize loading on those factors, with out-of-sample comparison against a marginal-elasticity design.","classification":"INDETERMINATE_RESEARCH_NEEDED","disposition":"REJECT","rationale":"The retrieved literatures cover nearly every component separately, but the shallow screen cannot establish whether the proposed closed-loop selection rule has already been tested. The hypothesis is also broad across fiscal and monetary instruments, so it should not advance without a focused review of one policy class and the bounded question."},{"hypothesis_id":"H5","search_queries":["margin liquidation systemic price impact optimization correlated portfolios","spectral systemic risk fire sales overlapping portfolios liquidation matrix","broker margin liquidation algorithm market impact cross asset solvency constraints"],"sources":[{"source_id":"H5-S1","title":"Fire sales, indirect contagion and systemic stress-testing","publisher":"Systemic Risk Centre, London School of Economics","url":"https://www.systemicrisk.ac.uk/sites/default/files/images/1.Eric%20Shaanning%20paper.pdf","source_class":"PRIMARY_RESEARCH","claims_supported":["Liquidity-weighted portfolio-overlap matrices are analyzed through eigenvalues and eigenvectors.","A few dominant eigenvalues can characterize the network with a low-dimensional factor model, and second-round spillovers can connect institutions that have no direct overlap."]},{"source_id":"H5-S2","title":"Liquidations","publisher":"Bulk Labs","url":"https://docs.bulk.trade/bulk-exchange/Liquidations","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","claims_supported":["The liquidation engine optimizes across an account's positions instead of closing only the largest loser.","It evaluates contribution to portfolio risk, preserves hedges, and liquidates only enough to restore margin health."]},{"source_id":"H5-S3","title":"Central clearing: trends and current issues","publisher":"Bank for International Settlements","url":"https://www.bis.org/publ/qtrpdf/r_qt1512g.pdf","source_class":"GOVERNMENT_OR_REGULATOR","claims_supported":["Margin tightening and collateral-value declines can force deleveraging and self-reinforcing fire sales.","Centralized exposure structures and margin practices can create endogenous system-level shocks."]}],"closest_analogue":"Portfolio-risk-optimized partial liquidation, combined with eigenvalue-based fire-sale stress testing of liquidity-weighted portfolio-overlap networks.","overlap":"Existing systems already optimize which positions to liquidate while preserving hedges and restoring solvency; research already uses dominant eigenmodes and low-dimensional representations to diagnose cross-account fire-sale amplification.","remaining_difference":"The testable difference is a live controller that coordinates or staggers liquidation instructions across separate accounts using the shared impact mode, while enforcing each account's solvency constraint and a latency-bounded reconstruction error. Existing retrieved implementations optimize within an account, while the spectral work is diagnostic rather than an execution engine.","classification":"POSSIBLE_DISTINCTION","disposition":"ADVANCE","rationale":"The component technologies are close and the space is crowded, but the systemwide cross-account control boundary remains materially different from both account-local liquidation optimizers and diagnostic fire-sale stress models. Replay under broker credit and latency constraints can resolve the distinction."}],"nominated_ids":["H1","H2","H5"],"replenishment_recommended":false}