{"schema_version":1,"assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","source_experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"invariant_mode_decomposition_design__economics_finance","archetype_slug":"invariant_mode_decomposition_design","domain_slug":"economics_finance","title":"Frozen modal warning and liquidity-stress controls for joint dealer deleveraging","opportunity_summary":"Evaluate whether a frozen invariant-mode model of dealer funding, balance-sheet, asset-sale, and market-liquidity data warns of synchronized deleveraging earlier than institution-level thresholds and a network stress test. Any later causal study would separately randomize liquidity buffers and liability terming, but the supplied endpoint can move mechanically because both assignments alter variables inside the modal coordinate and does not require improved downstream market liquidity.","adopter_authorizer":"The legally authorized financial supervisor, subject to independent model-risk and ethics review; participating firms must consent wherever participation is not compulsory.","scores":{"meaningful_impact":{"score":4,"rationale":"If the hypothesized hidden joint deleveraging direction exists, earlier recognition could reduce system-wide liquidity loss and avoid indiscriminate firm restrictions. The packet does not establish prevalence, frequency, or realized impact, preventing the highest score."},"stakeholder_pull":{"score":2,"rationale":"Supervisors, dealers, counterparties, clients, investors, and the public are identifiable affected parties, but the sealed record contains no evidence that any supervisor or dealer currently demands, funds, or would participate in this approach."},"incremental_advantage":{"score":3,"rationale":"The proposal defines sealed improvements of at least 0.05 in precision-recall area and five business days of lead time against two relevant comparators at the same alert budget. No observed result supports those gains, and the later causal endpoint can improve mechanically without downstream liquidity benefit."},"distinctiveness_plausibility":{"score":2,"rationale":"The composition of invariant modes, frozen alert testing, and randomized balance-sheet controls is specific and testable, but prior-art status is explicitly UNSEARCHED and no closed-book basis establishes distinctiveness."},"technical_implementability":{"score":3,"rationale":"The operator, data fields, data splits, modal gates, comparators, randomization, negative controls, and rejection rules are substantially specified. Implementability remains mixed because panel availability and measurement stability are unverified, dealer behavior is endogenous, and explicit tolerances for non-normality, pseudospectral sensitivity, transient growth, and mode swapping are absent."},"adoption_authority_feasibility":{"score":2,"rationale":"A potential authorizer and committee structure are named, with consent where required, but legal authority, confidential-data permissions, dealer participation, assignment feasibility, and the acceptability of changing 2% of assets or liabilities remain unverified."},"evidence_readiness":{"score":2,"rationale":"The candidate supplies a rigorous retrospective comparison and falsifiers, but it provides no observed data-quality result, sealed-period performance, intervention result, downstream outcome evidence, or independent feasibility finding."},"safety_net_benefit":{"score":4,"rationale":"The first step is nonoperative, live activity requires separate approval, and the packet specifies confidentiality restrictions, exclusions, placebo tests, subgroup financing-cost monitoring, halt conditions, and rollback. Residual risk remains from false reassurance, spillovers, procyclicality, and mechanically favorable causal endpoints."},"scalability":{"score":2,"rationale":"The method could in principle be repeated across a fixed panel, but the packet prohibits extrapolation beyond the declared panel or regime and requires confidential multi-firm data, custom validation, supervisory authority, drift monitoring, and coordinated firm-level interventions."}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"250K_TO_1M","scope":"Independent prior-art, legal-authority, data-completeness, measurement-stability, and trial-feasibility review, followed only on satisfactory findings by a frozen retrospective shadow evaluation against both declared comparators.","confidence":"LOW","assumptions":["Existing historical dealer-panel data can be accessed without purchasing or constructing a new multi-year dataset.","The work is limited to authorized analysts and does not include live alerts or balance-sheet controls.","Costs include quantitative, legal, data-governance, model-risk, partner-coordination, and evaluation labor."]},"initial_deployment_startup":{"band_2026_usd":"1M_TO_5M","scope":"If retrospective gates pass, establish secure recurring data pipelines, model governance, monitoring, audit controls, dealer agreements, randomization infrastructure, downstream outcome definitions, and independent trial oversight.","confidence":"LOW","assumptions":["Multiple dealers and supervisory teams must coordinate.","Existing supervisory technology can host part of the workflow.","The band excludes the recurring economic burden of holding added liquid assets or extending liability maturities."]},"operational_launch":{"band_2026_usd":"OVER_25M","scope":"Conduct the approved consenting-dealer stepped-wedge trial, including implementation and opportunity costs of the 2%-of-assets liquidity-buffer and liability-terming assignments, secure monitoring, subgroup-harm surveillance, and independent analysis.","confidence":"LOW","assumptions":["The trial includes multiple materially sized dealers.","Resource-equivalent cost counts balance-sheet funding and liquidity opportunity costs, not merely administrative spending.","Treatments are maintained long enough to test the five-business-day and next-month outcomes.","Dealer count, asset scale, funding spreads, and rollout duration are absent, so the selected band is highly assumption-sensitive."]},"annual_recurring":{"band_2026_usd":"5M_TO_25M","scope":"Operate a supervisory shadow-warning program after validation, including data ingestion, model re-estimation and drift checks, security, governance, analyst review, audits, and partner coordination; separately authorized balance-sheet controls are not assumed to be continuously imposed.","confidence":"LOW","assumptions":["The recurring program remains advisory or shadow-based rather than automatically imposing controls.","Several institutions provide confidential data on a regular schedule.","If balance-sheet controls become ongoing, recurring resource-equivalent cost could exceed this band."]}},"research_burden":"VERY_HIGH","earliest_credible_horizon":"12_TO_36_MONTHS","pipeline_gates":{"recognizable_externally_supportable_problem":{"status":"YES","reason":"The candidate states an observable systemic-liquidity problem independently of the modal solution, identifies affected actors and outcomes, and provides a falsifier based on failure to add sealed-period warning beyond both comparators."},"identifiable_adopter_or_authorizer":{"status":"YES","reason":"The legally authorized supervisor is explicitly identified as decision authority, with independent model-risk and ethics committees and firm consent where participation is not compulsory."},"distinct_testable_incremental_claim":{"status":"YES","reason":"At a matched false-alert rate and alert budget, the frozen model must improve final-block precision-recall area by at least 0.05 and median lead time by at least five business days versus both institution thresholds and a frozen network stress test."},"bounded_next_evidence_step":{"status":"YES","reason":"The packet authorizes an audit and a frozen retrospective shadow evaluation only, with predefined splits, comparators, modal acceptance thresholds, predictive thresholds, falsifiers, and no live control."},"no_unresolved_safety_or_authority_stop":{"status":"UNCERTAIN","reason":"The retrospective step is bounded and nonoperative, but confidential-data authority, legal authority, measurement stability, and trial feasibility are explicitly unverified. Any live trial also requires separate approval and consent where required."},"implementation_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The packet defines major activities and treatment magnitudes but supplies no dealer count, balance-sheet scale, data-access burden, rollout duration, funding-cost effect, or implementation resource evidence; only assumption-dependent broad bands can be assigned."}},"blocking_evidence":["Independent prior-art evidence establishing whether the warning design or mechanism composition offers a distinct incremental contribution.","Authorized access to a sufficiently complete, stable, time-indexed fixed-dealer panel with documented definitions and missingness.","Frozen non-normality, eigenvector-conditioning, pseudospectral-sensitivity, transient-growth, and mode-swapping rejection criteria.","Sealed final-block evidence that the modal alert clears both the 0.05 precision-recall-area and five-business-day lead-time gains against both comparators.","Independent legal, model-risk, ethics, confidentiality, consent, and dealer-participation feasibility findings.","A causal design requiring improvement in a sealed downstream market-liquidity outcome and distinguishing changed propagation from arithmetic movement caused by altering modal-coordinate inputs.","Evidence that assignments, anticipation, spillovers, adaptation, and subgroup financing-cost harms can be measured and controlled adequately."],"next_evidence_step":"Through an independent supervisory-data partner, complete a time-bounded prior-art, legal-authority, data-completeness, measurement-stability, and non-normality-diagnostic audit; only if those checks pass, run the frozen retrospective shadow evaluation on the declared split and alert budget, comparing the modal alert with both institution-level thresholds and the frozen network stress test. Stop if either comparator is not exceeded by at least 0.05 precision-recall area and five business days of median lead time, or if any modal stability, residual, placebo, or interpretation gate fails; do not deploy live alerts or controls.","research_questions":["Does the declared fixed-dealer panel exist with consistent definitions, adequate history, stable coverage, and legally authorized access?","Does one preregistered joint mode repeatably exceed both comparators on the unopened final block at the same false-alert rate and alert budget?","Do non-normality, ill-conditioned eigenvectors, pseudospectral sensitivity, transient growth, mode coupling, or mode swapping invalidate the invariant-mode interpretation?","Can a downstream market-liquidity endpoint be frozen so intervention success cannot be established merely by changing variables contained in the modal coordinate?","Can mediation analysis distinguish a genuine change in stress propagation from direct coordinate arithmetic?","Are concealed schedules, lack of pre-notice anticipation, stable measurement, and the preregistered interference mapping credible in a consenting-dealer stepped-wedge trial?","Do the liquidity-buffer and liability-terming assignments reduce systemic-liquidity harm without worsening liquidity elsewhere or increasing the worst affected client subgroup's financing cost by more than 10 basis points?","What legal authority, consent, confidentiality, model-governance, rollback, and cost-sharing arrangements would the authorizing supervisor and participating dealers accept?","What relevant prior methods or deployments already exist, and does this proposal retain a distinct testable advantage after that review?"] ,"recommendation":"PARTNERED_RESEARCH","uncertainty_constraints":["Closed-book assessment cannot establish prior art, novelty, prevalence, market size, stakeholder demand, data availability, legal feasibility, realized effectiveness, or exact cost.","The problem and lever remain hypotheses; the packet reports no empirical retrospective or randomized result.","Operational cost bands depend heavily on undisclosed dealer count, asset scale, funding spreads, data infrastructure, and trial duration.","The proposed intervention endpoint is vulnerable to mechanical movement because assigned variables enter the modal coordinate.","Strategic adaptation, anticipation, spillovers, nonlinear regime shifts, and non-normal transients may break the proposed interpretation.","Findings cannot be extrapolated beyond the declared dealer panel or regime."],"closed_book_prior_art_boundary":"Prior-art status is UNSEARCHED. No claim is made about novelty, prevalence, competing deployments, market size, or superiority outside the candidate's predefined comparisons; these require external research."}