{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"bounded_rivalry_governance__accounting_auditing:P3:v0","cell_id":"bounded_rivalry_governance__accounting_auditing","search_queries":["shared service cost allocation disputes business units allocation drivers survey managerial accounting","site:imanet.org shared services cost allocation drivers managerial accounting","cost allocation gaming business units allocation basis research","model risk management validation holdout testing accounting allocation model governance OCC Federal Reserve SR 11-7","academic study internal cost allocation incentives business unit managers manipulation cost allocations","site:pubsonline.informs.org cost allocation business unit managers incentives allocation bases","site:imanet.org managerial costing conceptual framework causality allocation cost drivers PDF","Technology Business Management standard cost allocation model cost pools towers cost drivers official","site:gao.gov shared services cost allocation business units audit allocation methodology","site:gov.uk shared services cost allocation methodology business units audit","site:gov technology business management cost allocation case study agency","site:tbmcouncil.org case study cost allocation business units chargeback transparency","site:imanet.org controller responsibility managerial accounting cost allocation","site:aicpa-cima.com controller cost allocation responsibility management accounting","corporate controller authority accounting policies cost allocation official guidance","site:gao.gov cost allocation basis reasonable consistent audit shared services"],"sources":[{"source_id":"S1","title":"Why do so many organisations struggle with cost allocation?","publisher":"Deloitte Switzerland","url":"https://www.deloitte.com/ch/en/services/financial-advisory/research/why-organisations-struggle-cost-allocation.html","source_class":"AUTHORITATIVE_SECONDARY","publication_date":"2021-04-04","accessed_at":"2026-08-03","claims_supported":["Large organizations allocate substantial corporate-function costs to business units.","Business units seek to protect profitability by avoiding corporate charges.","Organizations repeatedly revisit allocation models because of fairness disputes.","Greater model complexity increases effort and cost, supporting a bounded-complexity design."]},{"source_id":"S2","title":"Who should pay for support functions?","publisher":"McKinsey & Company","url":"https://www.mckinsey.com/capabilities/operations/our-insights/who-should-pay-for-support-functions","source_class":"AUTHORITATIVE_SECONDARY","publication_date":"2023","accessed_at":"2026-08-03","claims_supported":["Controllers and managers report opaque, poorly documented allocation systems.","Business-unit leaders express a need for actionable allocation information tied to controllable spending.","Allocations can charge units for costs their behavior cannot change.","Organizations should avoid unnecessary allocations and use them selectively where demand and cost are linked."]},{"source_id":"S3","title":"Distorted Cost Allocation: An Encouragement or Discouragement?","publisher":"Seoul National University Institute of Management Research / Seoul Journal of Business","url":"https://www.kci.go.kr/kciportal/landing/article.kci?arti_id=ART002659210","source_class":"PRIMARY_RESEARCH","publication_date":"2020","accessed_at":"2026-08-03","claims_supported":["Overhead allocation affects divisional incentives as well as decision information.","Observed over- and under-allocation were associated with lower subsequent divisional performance and lower firm performance.","Accurate cost drivers and distortion are consequential managerial-accounting concerns."]},{"source_id":"S4","title":"The Conceptual Framework for Managerial Costing","publisher":"Institute of Management Accountants","url":"https://www.imanet.org/-/media/df27d1327c4b4f9dbc0b5a43537bcc7f.ashx","source_class":"OFFICIAL_GUIDANCE","publication_date":"2017","accessed_at":"2026-08-03","claims_supported":["Managerial cost models should represent cause-and-effect relationships between resources and operations.","Useful models should connect monetary effects to operational causes and decision consequences.","Objectivity, accuracy, verifiability, measurability, materiality and data integration constrain model design.","Causal managerial costing is intended to support internal management decisions."]},{"source_id":"S5","title":"TBM Modeling Allocation Methods","publisher":"Technology Business Management Council","url":"https://www.tbmcouncil.org/learn-tbm/tbm-modeling/allocation-methods/","source_class":"STANDARD","publication_date":"2025","accessed_at":"2026-08-03","claims_supported":["Established practice allocates shared infrastructure and internal-service costs to business units using drivers.","Allocation rules are expected to be traceable, defensible, relevant, consistent and scalable.","Consumption, usage and other cost drivers are already standardized alternatives.","The standard provides model-design prior art but not a sponsor-rivalry tournament."]},{"source_id":"S6","title":"Supervisory Guidance on Model Risk Management","publisher":"Board of Governors of the Federal Reserve System","url":"https://www.federalreserve.gov/frrs/guidance/supervisory-guidance-on-model-risk-management.htm","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2026-04-17","accessed_at":"2026-08-03","claims_supported":["Model testing may include out-of-sample and out-of-time tests and comparisons of alternative assumptions and methodologies.","Sound model governance includes clear roles, controls, accountability and management of conflicts of interest.","Validation and ongoing monitoring are established practices.","Internal audit should assess governance effectiveness rather than duplicate model development or validation."]},{"source_id":"S7","title":"The TBM Taxonomy","publisher":"IBM Apptio","url":"https://www.ibm.com/docs/en/apptio-commercial/tbm-studio/saas?topic=fundementals-tbm-taxonomy","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"undated","accessed_at":"2026-08-03","claims_supported":["Commercial software already imports general-ledger data into cost pools and performs driver-based allocation.","Existing systems trace costs through resources and services to consuming business units.","User counts, transactions and application usage are established allocation inputs.","Software infrastructure for reproducible allocation models exists, although tournament governance is not documented."]},{"source_id":"S8","title":"HHS Policy for Information Technology Portfolio Management (PfM)","publisher":"U.S. Department of Health and Human Services","url":"https://www.hhs.gov/digital/governance/it-policy-archive/hhs-policy-information-technology-portfolio-management.html","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2021-09-23","accessed_at":"2026-08-03","claims_supported":["HHS is an identifiable institutional adopter of standardized technology cost modeling.","HHS requires its operating and staff divisions to use TBM when reporting IT costs.","The policy identifies CIO and governance-board authority, TBM analysts, authoritative data repositories and review cycles.","The policy expresses a need for accurate, detailed and transparent cost information but does not request a model tournament."]}],"problem_evidence":{"support":"STRONG","rationale":"Independent consulting evidence reports recurrent allocation disputes, opacity, substantial administrative effort and business-unit incentives to avoid charges. Primary research connects allocation distortion to divisional incentives and performance. These sources support the general problem, but they do not establish how often units strategically sponsor self-favoring models or coordinate exclusions—the proposal's more specific rivalry diagnosis.","source_ids":["S1","S2","S3"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Business-unit leaders and controllers express demand for actionable, transparent allocation information, and HHS provides a named adopter with explicit cost-modeling roles, authority and review processes. No identified organization expresses demand for identity-blinded tournaments, anti-collusion screens or annual competitive designation, so pull for the full intervention remains unverified.","source_ids":["S2","S8"]},"prior_art":{"proximity":"ADJACENT_PRIOR_ART","closest_analogues":[{"name":"TBM driver-based cost allocation","similarity":"Uses governed cost pools, consumption drivers, traceable allocation rules and business-unit chargeback/showback.","remaining_difference":"It standardizes and operates allocation models but does not stage rivalry among conflicted model sponsors using blinded holdouts, fouls, resource caps and expiring designation.","source_ids":["S5","S8"]},{"name":"IBM Apptio TBM Studio","similarity":"Provides ledger ingestion, driver-based allocations, cost-flow lineage and business-unit outputs needed to implement candidate models.","remaining_difference":"Product documentation describes model execution and reporting, not tournament entry, sponsor-conflict controls, comparative hidden testing or anti-coordination monitoring.","source_ids":["S7"]},{"name":"Interagency model-risk management practice","similarity":"Uses independent validation, out-of-time testing, alternative-method comparisons, conflict-aware roles and continuing monitoring.","remaining_difference":"It governs individual or portfolio model risk rather than distributive rivalry for a single internal cost-allocation standard.","source_ids":["S6"]},{"name":"IMA causal managerial-costing framework","similarity":"Evaluates cost models against causality, decision usefulness, verifiability, measurability and materiality.","remaining_difference":"It supplies evaluation principles but no competitive selection arena, blinded rank, sponsor-benefit test or temporary winner rule.","source_ids":["S4"]}],"distinctive_claim_remaining":"Conditional on direct tracing being infeasible and sponsors having distributive exposure, a preregistered, identity-blinded holdout comparison with independent re-performance will select an allocation model with better out-of-period stability, causal usage linkage and reproducibility than the incumbent controller-selected model or a non-blinded expert-panel selection, while reducing the association between sponsor benefit and selection rank within a capped resource envelope. This is falsified if rankings are holdout-sensitive, sponsor benefit remains the dominant rank predictor, no model clears reconciliation and auditability gates, or direct tracing is feasible at acceptable burden.","confidence":"MODERATE"},"implementation_evidence":{"support":"MODERATE","rationale":"Ledger-fed driver models, cost-flow lineage, alternative-method comparison, out-of-time testing, independent validation and governance roles are all implemented in existing standards, products or official practice. Evidence does not establish that sponsor identity can actually be blinded, that historical usage offers a valid prospective target, that anti-collusion screens work in this setting, or that a composite tournament score improves decisions.","source_ids":["S4","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Allocation distortion can affect unit performance signals and firm decisions, while large organizations report material allocation volumes and recurrent disputes. Magnitude for any particular pool is unmeasured.","source_ids":["S1","S2","S3"]},"stakeholder_pull":{"score":3,"rationale":"Stakeholders clearly want transparent, actionable allocations and at least one named institution mandates standardized cost modeling, but no adopter requests the tournament mechanism.","source_ids":["S2","S8"]},"incremental_advantage":{"score":2,"rationale":"The intervention plausibly adds conflict and rivalry controls to established costing and validation practice, but no comparative evidence shows improvement over independent controller judgment, causal-costing review or direct metering.","source_ids":["S4","S5","S6"]},"distinctiveness_plausibility":{"score":3,"rationale":"The searched standards, guidance and product documentation contain most technical components but no close match for the combined blinded, time-limited sponsor tournament. World novelty and patentability remain unmeasured.","source_ids":["S4","S5","S6","S7","S8"]},"technical_implementability":{"score":4,"rationale":"Existing products and standards support ledger reconciliation, driver models, lineage, business-unit outputs, alternative-model testing and validation. Blinding and valid prospective scoring remain difficult but do not block a retrospective shadow test.","source_ids":["S5","S6","S7"]},"adoption_authority_feasibility":{"score":4,"rationale":"A nonposting shadow exercise fits identifiable finance, CIO, analyst and governance roles, and HHS demonstrates that an organization can mandate cost-modeling methods and review cycles. Corporate controller authority at a prospective private adopter must still be confirmed.","source_ids":["S6","S8"]},"evidence_readiness":{"score":3,"rationale":"A closed-year comparison is well bounded and can use existing governed data, but the decisive evidence—real sponsor exposure, candidate submissions, holdout performance and direct-tracing feasibility—is organization-specific and proprietary.","source_ids":["S5","S6","S7"]},"safety_net_benefit":{"score":4,"rationale":"A nonposting retrospective shadow comparison preserves the approved allocation and can stop on reconciliation, authorization or independence failure. Risk rises sharply if rankings leak into budgets, compensation, tax or external reporting.","source_ids":["S6","S8"]},"scalability":{"score":3,"rationale":"Reusable rulebooks and commercial allocation infrastructure support replication, but driver validity, pool boundaries, data access, materiality and scoring weights must be tailored to each service pool.","source_ids":["S4","S5","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"Pre-registration plus reconstruction of the incumbent and up to three alternatives for one closed year and one pool, with one withheld quarter and a short independent re-performance review.","confidence":"LOW","assumptions":["Existing governed ledger and usage data are accessible without new procurement.","Roughly 3-8 staff-weeks across accounting, data and validation roles.","No live posting, participant compensation or external legal opinion."],"source_ids":["S1","S6","S7"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Design a reusable rulebook, governed extracts, reproducible model interface, conflict controls, audit logging, stress tests and appeal procedure for one organization.","confidence":"LOW","assumptions":["Existing allocation or analytics tooling can be configured rather than replaced.","One shared-service domain and a small number of models.","Internal validation and data-governance capacity are available."],"source_ids":["S5","S6","S7","S8"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Run the first controlled parallel-ledger year for one material pool, including model onboarding, independent validation, access controls, monitoring, stakeholder appeals and remediation capacity.","confidence":"LOW","assumptions":["No statutory or tax treatment changes.","Integration with several operational data sources is required.","External assurance or specialized engineering may be needed.","The band excludes the shared-service costs being allocated."],"source_ids":["S5","S6","S7","S8"]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Annual challenge cycle for one pool: data refresh, model maintenance, holdout and stress testing, independent re-performance, appeals, monitoring and ex-post review.","confidence":"LOW","assumptions":["Three or fewer challengers per cycle.","Existing infrastructure and policy artifacts are reused.","Material data-model changes or litigation are excluded."],"source_ids":["S1","S6","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Multiple independent publishers and primary research support allocation disputes, opacity, distorted incentives and consequential decision effects.","source_ids":["S1","S2","S3"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"HHS is an identifiable adopter that mandates TBM cost reporting, names CIO and governance authority, and assigns analysts and stakeholders to operate and review cost models. This verifies a credible adopter/authorizer structure, not demand for the proposed tournament.","source_ids":["S8"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The proposal can compare blinded tournament selection with the incumbent controller model and non-blinded expert selection on preregistered stability, causal-linkage, reconciliation, reproducibility, burden and sponsor-benefit outcomes.","source_ids":["S4","S6"]},"bounded_next_evidence_step":{"status":"YES","reason":"One closed-year, one-pool, nonposting shadow tournament with at most three alternatives is bounded, reversible and has explicit comparators and stopping rules.","source_ids":["S6","S7"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"For the shadow step only, existing governance roles and data permissions can authorize retrospective analysis while prohibiting ledger, budget, compensation, tax and external-reporting use. Live adoption would require a new authority review.","source_ids":["S6","S8"]},"credible_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The scopes are explicit resource-equivalent estimates and existing sources identify relevant work and roles, but no direct pricing or observed implementation-cost evidence was found for this novel tournament configuration.","source_ids":["S1","S6","S7","S8"]}},"next_evidence_step":"With a named controller and data owners, preregister a closed-year shadow study for one reconciled shared-service pool. Compare (A) the incumbent controller-selected basis, (B) a conventional non-blinded independent accounting-panel choice, and (C) the proposed blinded holdout tournament using the same candidate models. Freeze eligibility, holdout quarter, weights and stress scenarios before outputs are inspected. Measure reconciliation failures, independent reproducibility, out-of-period stability, sensitivity to discretionary assumptions, causal usage linkage, data and labor burden, rank correlation with sponsor savings, unexplained transfers to nonparticipants and direct-tracing feasibility. Do not post results or use them for budgets, compensation, tax, transfer pricing or external reporting. Stop if common pool reconciliation, authorized data access, validator independence or prospective target validity fails. Do not proceed toward live adoption unless C is reproducible and materially improves at least one preregistered decision-quality measure without worsening the others or increasing sponsor-benefit association.","blocking_evidence":["No organization-specific evidence shows that sponsors systematically submit self-favoring models or influence selection after legitimate service heterogeneity is controlled.","No comparative study shows that blinded tournament selection outperforms controller judgment, an independent expert panel or direct tracing.","The validity and temporal stability of candidate usage drivers require proprietary operational data and live or retrospective testing.","Sponsor identity may be inferable from model structure, defeating blinding.","The composite score's weights and ground-truth usage-linkage measure are unresolved value judgments.","Actual implementation and recurring costs have not been observed.","Corporate authority and legal boundaries for any live use in budgeting, compensation, tax, transfer pricing or external reporting remain organization-specific."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search found established causal-costing guidance, TBM standards, commercial driver-allocation systems, institutional cost-model governance and official out-of-time model-validation practice, but no direct source describing the full identity-blinded, sponsor-conflict-bounded, anti-coordination, resource-capped and annually expiring allocation-model tournament. This bounded search does not measure world novelty, patentability, freedom to operate, market size or realized impact.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Recruit a named controller-authorizer, independent validator and governed-data owner for a nonposting study.","Establish from proprietary records whether sponsor exposure predicts proposed allocation shares, exclusions, window choices or selection influence after controlling for service heterogeneity.","Preregister the incumbent and non-blinded-panel comparators, holdouts, scoring weights, minimum meaningful differences and falsifiers.","Demonstrate independent reconstruction and stable rankings across reasonable holdouts and stress scenarios.","Determine whether direct tracing or pool decomposition is feasible at lower authorized burden.","Record actual staff time, tooling, assurance and governance costs to replace the low-confidence cost bands.","Complete a separate accounting-policy, tax, transfer-pricing, privacy and personnel-use review before any live trial."],"reason":"Open-web evidence verifies the broad allocation problem, stakeholder need, credible governance roles and adjacent technical practice, but cannot establish the candidate's decisive organization-specific premises or comparative effect. Those require proprietary data and a controlled shadow test. Under the controller rule, that evidence dependency requires STOP_EMPIRICAL_RESEARCH_NEEDED, and every STOP is non-repairable in this evaluation cycle."},"proposal_index":3}