RESEARCH DOSSIER
Claims Expense Attribution and Settlement Cost Accounting
Abstract. Claims expense attribution is the measurement and assignment of the resources consumed in investigating, administering, defending, resolving, and paying insurance claims. Settlement cost accounting uses that attribution to describe the economic consequences of a claim process across transactions, service providers, administrative stages, and resolution pathways. This research dossier examines the distinction between claimant compensation and the expense of determining and delivering it; the relation between allocated and unallocated loss adjustment expenses and US statutory reporting categories; activity-based and time-driven cost allocation; internal labor, outsourced services, litigation and recovery expenses; expected pathway costs; ledger and actuarial reconciliation; and governance of operational incentives. Its methodology combines a purposive documentary review of authoritative professional, supervisory, and accounting texts, selected actuarial literature and empirical research summaries, and reproducible synthetic calculations. The principal finding is that cost attribution requires the joint specification of a cost object, resource boundary, time basis, and allocation rationale. An assigned cost can faithfully represent resource consumption without representing incremental avoidable expenditure or an insurance contract liability. Equally, lower observable settlement costs do not establish superior claim outcomes when cases differ in complexity, legal obligations, timing, or service quality. The study demonstrates these distinctions through a capacity-constrained costing example, competing settlement pathways, an expense rollforward, and portfolio sensitivity analyses. It provides analytical synthesis within established disciplines; it introduces no new scientific theory and makes no claim of original empirical discovery.
Keywords: Claims science; loss adjustment expense; allocated loss adjustment expense; unallocated loss adjustment expense; claims handling; settlement costs; activity-based costing; practical capacity; defense and cost containment; adjusting and other expenses; vendor management; insurance accounting; expense reconciliation; insurance conduct; cost attribution; operational control.
Introduction
Insurance institutions expend resources both to meet covered obligations and to establish what those obligations require. Claims administration generates an independent stream of resource consumption even when investigation establishes that no payment is owed. An insurer must record the consequence of the insured event, the adjudication of the claim, and the work required to deliver the resulting decision. The cost of the third activity is the central object of this dossier.
The parent research field, Claims Science, treats the claim as a legal, economic, evidentiary, operational and financial relation. This study concentrates on the operational and managerial consequences of the work performed in that relation. Detailed insurance contract liability measurement, reserve estimation, and the law of compensation belong to their respective research domains. Their interface with operational costing is studied insofar as it affects classification, reconciliation or interpretation.
Three research questions structure the study. Which observable activities provide defensible drivers for internal and external expenses? How do heterogeneous claim pathways change the composition, amount and timing of resource consumption? How can operational allocations reconcile with accounting and actuarial information while retaining their distinct purposes?
The research method uses the relevant text of named documents or the expressly identified accessible portion of a study. Searches combined the terms claims adjustment expense, unallocated loss adjustment expense, activity-based costing, insurance claims accounting, settlement litigation, and claims handling regulation, with searches on the standard setters' and publishers' domains. Materials were selected for direct relevance, identifiable authorship or authority, clear jurisdiction or population, and accessible supporting content. Publisher abstracts support only their explicit statements. Sources that could not be substantiated were excluded from substantive claims. This is a targeted documentary synthesis, not an exhaustive systematic review or meta-analysis. The evidence record identifies the checked passages and source boundaries.
Every worked example below is synthetic, denominated in US dollars for convenience, and uses explicitly stated simplified assumptions. No private case files, claimant information, proprietary insurer ledger or unpublished vendor data were obtained. Examples test internal consistency, interpretation and edge cases; their numerical parameters are not industry estimates. The international scope is comparative, with examples from US actuarial and statutory practice, the UK conduct framework, a New Zealand implementation of international accounting requirements and the IAIS supervisory framework. The analysis separates general managerial propositions from jurisdictional rules.
Claims Expense Categories and Units
The cost of the claim and the cost of claims work
A claim payment transfers financial value in fulfillment of an insurance obligation or settlement. Claims work consumes time, technology, professional services, and organizational capacity to investigate, decide and execute that obligation. A repair invoice paid to restore insured property may constitute a covered benefit; an engineering invoice for determining the cause and extent of damage may constitute adjustment expense. Their recipient may even be the same enterprise. Classification requires the economic purpose and contractual arrangement of each invoice line.
The US Actuarial Standards Board (2007) treats the administration, investigation of coverage, settlement and defense of claims as adjustment expenses under ASOP No. 43, including work on claims ultimately found invalid. The distinction explains why closed-without-payment files must remain in a processing cost denominator. An expense ratio using indemnity as its only denominator cannot describe a portfolio of investigated but unsuccessful demands.
In traditional managerial and actuarial terminology, allocated loss adjustment expenses (ALAE) are claim-linked expenses, commonly identifiable with particular files; unallocated loss adjustment expenses (ULAE) are expenses recorded at a shared claims-function level. Organizational definitions and historical reporting conventions vary. The ability to assign a transaction to an individual case is a property of the available records and the allocation policy. It can change when time recording or invoice coding improves. A reclassification from shared expense to assigned case cost may reflect better measurement without any change in the total organizational expenditure.
A useful cost classification therefore retains four dimensions. Traceability concerns whether a resource or transaction is identifiable with a cost object. Function concerns what service the expenditure supported. Behavior concerns whether expenditure changes when activity changes over a specified horizon. Financial eligibility concerns treatment under an identified reporting framework. These attributes answer different questions; forcing them into one mutually exclusive code produces category mistakes.
DCC and A&O
The NAIC (2025) Property/Casualty Annual Statement Blank presents separate lines for defense and cost containment (DCC) and adjusting and other (A&O) expense. These reporting dimensions classify expense according to statutory categories. ALAE and ULAE, as used in this operational study, distinguish traceability rather than statutory function. A specific independent adjuster fee may have a case identifier without being a defense cost. A defense-related internal service may be shared across files. A crosswalk must thus use the applicable annual statement instructions and expense definitions rather than assume that every ALAE item equals DCC or every ULAE item equals A&O.
For example, an invoice for counsel appearing for an insured in litigation, an invoice for a coverage opinion, and the payroll allocation for a claims lawyer may all involve legal expertise. The type of legal work and its purpose matter more than the profession named in the vendor master. An operational ledger should preserve a detailed legal activity code and distinguish defense of the insured from analysis of the insurer's coverage position. A statutory mapping can then be applied and reviewed separately. This dossier does not provide an authoritative DCC/A&O classification for every individual statutory expense type.
HM Revenue & Customs' general insurance glossary provides a UK terminology illustration that includes both specific adjustment services and associated administrative overhead. It is a descriptive glossary, rather than current financial-reporting or conduct legislation. The common analytical feature is the inclusion of the resources consumed in administering the claim, whether an external payment or a share of an internal function.
Cost objects and time bases
A claim file is often the smallest administrative object, but it can contain multiple coverage features, claimants, beneficiaries, vehicles or allegations. A claim event can generate hundreds of files, and a single litigation engagement can support a group of related proceedings. A transaction can serve more than one claim over several reporting periods. The architecture should preserve distinct event, claim, coverage, activity, vendor engagement, invoice line, payment and reporting-period identifiers.
The cost object must match the inquiry. Claim-level allocation supports investigation of service intensity. The coverage feature supports comparisons where separate bodily injury and property damage work are performed. Event-level allocation can identify catastrophe response resource requirements. Vendor engagement supports procurement and budget oversight. A cohort or accounting group supports aggregation, but aggregate financial measurements require additional assumptions.
Dates also answer different questions: service date, invoice receipt, accrual recognition, payment, loss occurrence, report, closure and reopening. A claim that closes in March may have February services invoiced in April and paid in May. A closure-cohort view assigns those expenses to the originating case; a calendar-period view identifies expense incurred during the relevant service period; a cash view identifies disbursement. The insurer should retain the three separate clocks and publish each denominator and measurement cutoff.
Full, incremental, marginal and opportunity cost
Full attributed cost includes directly traced services and an allocated portion of supplied shared capacity. An incremental cost is a change in total expenditure associated with a specified decision over a specified horizon. Marginal cost is the incremental change associated with a small change in activity. Opportunity cost is the value of an alternative use of a scarce resource foregone because the resource is committed elsewhere.
An employee with a fixed salary may consume two hours on a claim; allocating salary cost measures the resource demand. Eliminating the activity may free two hours while leaving current payroll unchanged. If backlog exists, the released capacity can have substantial operational value. If the capacity cannot be redeployed, the cash saving in the current month may be near zero. A multi-year staffing decision creates a different incremental cost calculation. Accordingly a full cost estimate is a description of consumed capacity, not automatically an estimate of cash avoided.
A common error is to compare an external invoice with the internal hourly wage alone. Internal delivery may also require management, technology, evidence retention, training and support; external delivery adds contract administration and retained oversight. A decision model should identify the costs that will differ among options, including constraints on existing capacity, while retrospective full costing should retain all resources actually supplied within its defined boundary.
Activity Cost Drivers
Cost pools and reliable drivers
Activity-based costing assigns resource costs to activities and activities to cost objects using identified driver quantities. A resource pool is a collection of expenses with sufficiently similar behavior and service characteristics to have a meaningful allocation denominator. A cost driver is a measurable approximation of the activity demand consuming that pool. The driver need not be perfect; it must be sufficiently stable, economically interpretable and testable for its decision purpose.
A claims department can distinguish notification and validation, coverage inquiry, document acquisition, technical inspection, valuation, communication, payment processing, litigation management, recovery, quality checks, and routine file maintenance. A driver taxonomy can include transactions, labor minutes, inspections, specialist reports, payment instructions, open claim-months and separately tracked exceptional tasks.
A uniform allocation by claim count assumes equal resource demand. This is defensible for an activity whose service requirement is approximately uniform. Allocating inspection costs by count regardless of inspection duration, or allocating litigation work across all claims regardless of whether they involve litigation, can systematically misstate complex file costs. A driver based on indemnity amount is equally fragile: severe losses may require lengthy investigation, but payment value may change without any change in processing demand.
Where measurements are possible, direct activity evidence should precede proxy allocation. For shared services with heterogeneous beneficiaries, a reasonable proxy may be necessary. Its selection should be documented with the resource relationship, data coverage, periodic review, and evidence of improvement over alternatives. A driver adopted for budget convenience should be identified as an allocation convention rather than a demonstrated causal relation.
Time-driven activity-based costing
Kaplan and Anderson's approach is described in Kaplan's (2005) published excerpt, Rethinking Activity-Based Costing. It relates a resource capacity cost rate to the estimated time consumed by an activity. For resource pool r in period t, let C(r,t) be the supplied resource cost and K(r,t) its practical capacity in minutes. Define the rate:
rate(r,t) = C(r,t) / K(r,t).
If q(i,r,t) is the observed or estimated number of consumed minutes for claim i, its attributed resource cost is rate(r,t) multiplied by q(i,r,t). Multiple pools can be combined by summation. Direct invoices are added separately only if they have been excluded from the shared pool. Practical capacity must have a documented definition: paid hours less clearly defined unavailable time under normal operating conditions. Training or quality work can instead be modeled as separate activities if included in the capacity denominator; it must not be excluded from available time and then counted again as a separate consumed minute without a corresponding adjustment.
An illustrative complexity equation for an adjuster's minutes is:
T = 20 + 15D + 45I + 30R,
where D indicates supplementary documentation, I indicates a site inspection, and R indicates a specialist review. A simple file with none uses 20 minutes; a file with all three uses 110. Repeated reviews should use a count rather than a binary indicator when repeats materially change time. The coefficients are hypothetical, not fitted observations. Differences across languages, regions, accessibility requirements or claim types must be checked before transporting a time equation to another operating environment.
Time drives resource assignment only where resources are substantially consumed in proportion to time. Per-transaction technology contracts may require transaction counts. A dedicated expert report may have a fixed contracted fee. A surge-capacity contract may contain both a readiness retainer and a per-deployment charge. Combining these under one rate sacrifices information about the expenditure behavior.
Controlled example of practical capacity
Assume a monthly adjustment team costs $120,000 and provides 80,000 practical minutes. Separately, a common payment platform costs $12,000 and can process 6,000 ordinary payment instructions during the period. The labor rate is $1.50 per minute; the platform rate is $2 per instruction. Three mutually exclusive activity cohorts consume the following resources:
| Cohort | Files | Labor minutes per file | Total minutes | Payment instructions | Direct vendor fees |
|---|---|---|---|---|---|
| Routine | 1,000 | 25 | 25,000 | 1,000 | $0 |
| Complex | 250 | 120 | 30,000 | 500 | $50,000 |
| No indemnity | 200 | 50 | 10,000 | 0 | $4,000 |
| Total | 1,450 | 65,000 | 1,500 | $54,000 |
Labor assigned to routine files is $37,500, complex files $45,000, and no-indemnity files $15,000. Payment platform allocation is $2,000, $1,000 and zero respectively. Including the separately traced vendor fees, the case-assigned totals are $39,500, $96,000 and $19,000. Across all files, assigned cost equals $154,500.
The labor pool has 15,000 unused practical minutes, worth $22,500 at the standard rate. The platform has 4,500 unused instructions, worth $9,000. The total supplied cost is $186,000, comprising $120,000 labor, $12,000 technology and $54,000 external fees. The identity is:
$154,500 + $22,500 + $9,000 = $186,000.
Mean assigned cost is $39.50 for routine files, $384 for complex files and $95 for no-indemnity files. The portfolio mean is about $106.55 per file while the supplied cost divided by file count is about $128.28. The difference is not missing expense; it is unused capacity not attributed to actual claims.
Under a fully absorbed labor allocation based on actual minutes, the labor rate becomes approximately $1.84615 per minute. The identical 120-minute complex file receives $221.54 of labor cost instead of $180. The apparent increase is driven by the smaller denominator, not more complex work. Both reports can be useful if their cost boundaries are declared. Actual-usage attribution is informative about resource demand; fully absorbed allocation is a convention for distributing total cost and conceals the explicit unused-capacity balance.
Estimation and validation of times
System timestamps record events, not necessarily active work. The time between document upload and review includes waiting. A claim open for fifty days may require only hours of professional attention. Assigning fifty days of salaried labor to it would confound duration with work intensity. Conversely, a stalled file can consume recurring maintenance, communication, storage, and oversight; open-time can be a driver for those specific services.
A credible observation program samples transactions stratified by task, complexity, staff grade, product, service mode and period. It records time definitions and protects against incentives to inflate or suppress observed work. Task concurrency, interruptions, travel, batching, supervision and unrecorded messages need explicit treatment. Time captured from interviews is economical but should be compared with observed activity or independently recorded work where material.
The estimator should distinguish observation error from actual heterogeneity. Averages alone conceal heavily skewed tasks. Reporting the median, upper quantiles and sample sizes within task types reveals whether a single standard time is suitable. Persistent residuals for a particular work type suggest that the time equation or classification is incomplete. A new workflow, vendor change or claimant service policy requires versioned recalibration. Previous reporting periods should retain their original rates or be clearly restated under the new basis.
Adjuster and Vendor Expense Attribution
Internal resources and organizational structure
Internal adjusters may perform administration, investigation, technical valuation, negotiation and post-settlement tasks within the same role. A salary recorded in one cost center cannot distinguish those activities. The managerial accounting process should preserve the employment cost pool and the assignment model. Compensation, employer contributions, benefits and attributable departmental support may be included according to the defined cost boundary; internal transfers should be eliminated where they duplicate supplied resources.
A centralized claims function often serves different insurance lines and legal entities. Separating shared costs from intercompany service charges is necessary for consolidated reporting. For entity-specific reporting, the service-charge reconciliation should identify the legal entity that incurred the expense, the receiving entity, the applicable agreement, the allocated activity and any consolidation elimination. The underlying service and the transfer-price charge should remain traceable even if the group eliminates the internal transaction.
Claims management may also involve staff outside the claims department, such as finance, IT, medical review and legal affairs. Department membership alone neither establishes nor excludes a claims function. A quarterly payroll total should be separated according to activity or directly attributable services. General corporate work should remain distinguishable from claim-specific or claims-function resource consumption rather than be spread blindly across closed files.
External services and invoice-level classification
External vendors may be engaged for field inspection, specialist appraisal, translation, forensic accounting, medical examinations, defense, mediation, payment processing, call center capacity, repair networks or recovery. An invoice identifier is an accounting document; its individual lines may support different claims, activities and contractual payment obligations. Service codes and engagement identifiers should be checked against the vendor contract, work authorization, delivered evidence and economic purpose of the service.
For a hypothetical $18,000 engineering invoice serving three claims, direct time records identify 30, 45 and 15 hours of equally priced work. The allocation is $6,000, $9,000 and $3,000. If $3,000 of the invoice is a separate event-level mobilization cost, the remaining $15,000 is allocated by those hours: $5,000, $7,500 and $2,500. The $3,000 remains in the event pool until a documented event-level driver is selected. A single invoice cannot justify a claim-level causal allocation when the supporting work record is absent.
Recurring vendor charges need decomposition before allocation. For a $10,000 retainer plus $100 per completed inspection, a month with forty inspections costs $14,000. The $100 marginal contractual price does not describe the entire service cost. Dividing $14,000 by forty gives $350 per inspection on a fully absorbed basis. If the retainer secures thirty days of catastrophe-response readiness, it can be analytically more useful to report $10,000 capacity and $4,000 consumed transactional services separately.
Contracts may contain minimum payments, stepped fees, milestone billing, cancellation provisions and reimbursement of reasonable disbursements. A service cost forecast must use these actual terms. A ratio of current vendor spend to completed claims can be distorted because invoices are received late or because a supplier was engaged in a previous-period case. Cutoff controls and accruals are thus central to evaluating external service performance.
Outsourcing, allocation and incentives
The NAIC (2026) annual statement expense instructions illustrate a reporting distinction between affiliated management services and non-affiliated claims-handling services, and require consistent allocation treatment in the relevant exhibit. These are US statutory reporting instructions, not a universal cost-accounting rule. The threshold and detailed reporting treatment should be read from the applicable filing-year document and jurisdiction; this study does not apply those filing thresholds to hypothetical cases.
A supplier can lower a quoted fee while increasing internal exception handling. A vendor whose performance is measured by closures can reduce recorded expense per closed file by prematurely marking cases complete, increasing reopenings and transferring work to another team. Conversely, a specialist may increase the expense of the first investigation while reducing repeated inspections, avoidable disputes, or incorrect payments.
A useful vendor comparison therefore begins at engagement approval and follows the same service horizon. Compare matched claim types, scope of deliverables, invoice completeness, lead times, subsequent work, quality and complaints. A new supplier used only for catastrophic or legally complex claims cannot be assessed by an unadjusted comparison with a supplier serving routine files. Selection into a supplier is part of the analysis.
Technology and digital claims handling
Digital systems can convert manually performed tasks into transaction charges, fixed platform commitments, and supervision or exception work. An automated payment validation that saves five active labor minutes per case may require interface maintenance, data-quality review, testing and fallback capacity. The managerial calculation compares expected changes in total resource consumption and expenditure over time. Reassigning staff without reducing their salaries still releases capacity; treating that release as immediate cash saved overstates the financial result.
Where a platform serves underwriting, claims and servicing, its cost pool should be split using measured resource demand or a defensible service proxy. Software subscription charges may be allocated by attributable user seats or processing units; cloud computing may be allocated by workload tags; shared development may require a documented lifecycle cost policy. Financial capitalization and amortization are separate from the operational cost assignment and must follow applicable accounting requirements.
Exception rates deserve special attention. An automated route operating on ninety-five routine claims while five complex exceptions require intensive manual review may be valuable, but the five exceptions cannot be costed as though they received the routine service. Changes in error detection, claimant access and escalation capacity should be measured alongside any reported cost improvement.
Settlement Pathway Cost Models
Pathway definition and horizon
Settlement paths are sequences of investigation and resolution activities conditional on the legal and factual circumstances. Examples include straightforward payment, technical assessment and negotiation, internal review, mediation, arbitration, litigation, dismissal and recovery. Many claims traverse several stages; an initially negotiated case may become litigated and then settle. A single final-path label is useful for descriptive reporting but loses the cost and timing of earlier stages.
A pathway cost includes only the expense boundary and horizon explicitly specified. A model may forecast avoidable future expense from today's decision, while a retrospective assessment includes the cost already incurred. Sunk cost should be recorded in the total case history but excluded from prospective comparisons when it is invariant to the decision. Expected claimant compensation must remain an independent component, with lawful entitlement and verified evidence governing settlement decisions.
The income and balance-sheet accounting treatment of expense is a further question. The following calculations are operational decision-support examples. They do not prescribe an offer amount, determine coverage, set a reserve or imply that financial cost can override a contractual or regulatory obligation.
Expected cost under uncertainty
For feasible pathway j, let scenarios s have conditional probabilities p(j,s) summing to one. Let I(j,s) be future indemnity, E(j,s) the future adjustment expenditure and R(j,s) the future recovery within the selected boundary. Expected undiscounted net outflow is:
V(j) = sum over s of p(j,s) times [I(j,s) + E(j,s) - R(j,s)].
Gross indemnity and expense, recoveries and timing should also be shown separately. Expected values can hide low-probability severe outcomes. A risk-neutral expectation is a descriptive summary, not a sufficient decision rule when tail exposure, claimant interests, legal duties, financial constraints or uncertainty about probabilities are material. Alternative feasible pathways should be assessed on these dimensions separately rather than through an arbitrary monetary penalty for claimant harm.
Suppose two pathways are already lawful alternatives for resolving a genuinely disputed hypothetical loss, all past expense is excluded, and no recovery is expected. Pathway A is a successful negotiated resolution with $20,000 future indemnity and $2,000 future handle-and-settle expense. Pathway B proceeds to formal dispute resolution with the following hypothetical outcomes:
| Outcome | Conditional probability | Indemnity | Adjustment expense | Total |
|---|---|---|---|---|
| Early agreement | 0.50 | $19,000 | $4,000 | $23,000 |
| Later determination | 0.35 | $24,000 | $10,000 | $34,000 |
| No indemnity after determination | 0.15 | $0 | $8,000 | $8,000 |
Pathway B expected indemnity equals $17,900, expected adjustment expense equals $6,700, and expected total is $24,600. Pathway A has total $22,000. The illustrative expected difference is $2,600 in favor of A. Its $20,000 indemnity is not justified by this comparison alone; the underlying claim merits and feasible lawful terms establish the decision space. B's cost distribution has an upper scenario of $34,000 and a lower scenario of $8,000, while A is modeled as certain by assumption. If an agreement is uncertain, A must also be modeled with a failure branch.
If 10 percentage points of B's probability move from later determination to no indemnity, its expected total falls by 0.10 times ($34,000 - $8,000), or $2,600, to $22,000. This break-even calculation shows the sensitivity to a plausible classification of case outcomes; it is not a prediction of litigation success. If counsel has already billed $5,000 in both routes, adding that sunk amount yields $27,000 and $29,600 in retrospective expected totals but leaves the prospective difference unchanged.
Time, delays and reopening
Paying an agreed settlement may close financial exposure, but the file can remain open for invoice settlement or recovery. Closure must be defined by its purpose. An administrative closure that is followed by a reopened claim incurs additional work; a payment that legally completes the insurer's obligation may precede administrative closure. Reporting cost per final closure requires a rule for reopened episodes and a consistent observation horizon.
Suppose 100 routine claims each have $50 of initial processing cost. Five reopen and incur $240 each. Initial processing cost totals $5,000; reopening adds $1,200. Cost per original file becomes $62. Counting each of the five second closures as a new claim gives 105 closures and $59.05 per closure, concealing some of the additional process burden. Reporting both original-file cost and episode-based cost clarifies the difference.
Duration can be represented as a driver of recurring work rather than substituted for active work. A claim requiring ten minutes of maintenance per open month costs $15 per month at the earlier $1.50 per minute capacity rate. A six-month difference creates $90 of modeled resource demand under the specified assumptions. It also extends the period over which any entitlement remains unresolved. An insurer cannot infer that six months of delay is permissible from a low measured maintenance cost.
Empirical evidence and pathway selection
Asmat and Tennyson's (2016) litigated auto claims study reports in its public abstract that changing state bad-faith tort liability affected observed dispute patterns in historical US automobile claims, with evidence consistent with differing expectations about trial outcomes. The result is a warning against transferring a fixed litigation probability across legal regimes or periods. The abstract does not establish the operational expense rates used in this monograph.
The Insurance Research Council (2026) auto injury claims summary reports greater duration among attorney-represented bodily injury claims in its studied population. It also documents increasing representation and litigation in the reported period. Representation and complexity may be jointly determined; a descriptive difference in closure duration does not identify the causal effect of representation. Claimant attorney fees are distinct from the insurer's defense expenditures and from the indemnity the insurer pays. The underlying report microdata were not examined.
An observational comparison of pathways should adjust for information available before assignment: line, coverage, injury or damage characteristics, jurisdiction, initial complexity, limits, reporting lag, and operational capacity. Variables measured after assignment, such as total litigation invoices, can be consequences of that pathway; conditioning on them may distort causal estimation. Missing factors such as the strength of evidence and genuine legal uncertainty can remain important even in rich data.
Matched or stratified comparisons can describe more comparable populations but require explicit overlap diagnostics. More credible causal estimates would require a plausible assignment mechanism or a research design with defensible assumptions, accompanied by sensitivity analysis. A case triage system should never impose experimentally inferior service merely to obtain a clean cost comparison. Ethical and legal constraints on research and operational decisions precede model selection.
Reconciliation to Financial Statements
The expense bridge
Operational expense attribution begins with a defined general ledger perimeter, maps transactions to activities, and reconciles allocated totals with source accounting balances. Each view can be independently correct while using a different time basis. The reconciliation should separate expenses recognized in the period, actual payments, opening and closing invoices payable, estimates of unbilled services, and projections of future not-yet-performed claims work.
For a simple calendar-period account with no foreign exchange, transfers or corrections, expense incurred for performed services is:
service expense = cash paid + closing payable - opening payable.
If $80,000 was paid, opening payable was $20,000 and closing payable was $35,000, incurred service expense is $95,000. If accrued but unbilled current services are included in the closing payable, they must not be added a second time. Suppose an independently estimated $150,000 of future adjustment work remains on previous loss events. This projection is not part of the $95,000 expense for performed services, and a bridge to financial recognition must consider the applicable measurement policy and any prior estimates already booked.
The synthetic monthly capacity calculation above produced $186,000 supplied expense. That amount may be reconciled to the transaction ledger as $120,000 labor, $12,000 platform, and $54,000 external fees. The activity subledger accounts for $154,500 assigned and $31,500 unconsumed capacity. If the finance team receives an additional $4,000 invoice for past-period services, its accounting recognition, attribution to original claim cohorts and cutoff treatment require separate consideration. A retrospective change to claim economics need not imply a current-month change in activity.
An operational accounting table should preserve the movement category: new original transaction, estimate, invoice matching, payment, credit note, accrual reversal, correction, allocation, transfer, and recovery. A reversal should carry the original transaction identity and sign. Deleting the original transaction removes the audit trail and can make corrected summaries incomparable.
IFRS 17 and the attribution boundary
The IFRS Foundation's IFRS 17 key terms describes fulfillment cash flows as expected insurance-related receipts and payments adjusted for timing and risk, and identifies the liability for incurred claims with past services. This educational explanation is read alongside the standard rather than treated as its normative wording.
The External Reporting Board's NZ IFRS 17 reproduces the relevant international accounting requirements for New Zealand reporting. Paragraph B65(f) includes the costs of investigating, processing and resolving claims under existing contracts. B65(l) includes directly attributable fixed and variable overhead, allocated consistently to groups of contracts using a rational systematic basis. B66 excludes, among other items, nonattributable costs, abnormal wasted resources and cash flows from reinsurance contracts held. The normative application depends on the applicable jurisdiction, reporting entity and version of the standard.
The traceability of a specific expense is therefore neither necessary nor sufficient for insurance-contract measurement eligibility. Internal claims support shared across a portfolio may qualify as attributable under the applicable standard; a transaction charged to an individual case might require separate treatment if it is outside the permitted boundary. An operational assignment may track the cost of errors and rework to identify process improvements while financial accounting distinguishes abnormal waste as required.
Consider a hypothetical $200,000 general ledger cost pool: $130,000 directly attributable routine claims processing and external services, $30,000 candidate attributable common overhead, $25,000 nonattributable corporate work and $15,000 identified abnormal waste. A managerial view could analyze $175,000 of claims-related spending, including the waste: $130,000 + $30,000 + $15,000. The remaining $25,000 is corporate work. A hypothetical candidate directly attributable perimeter is $160,000, conditional on the actual accounting assessment. The $40,000 remaining consists of $25,000 corporate work and $15,000 abnormal waste. No amount is automatically a liability; the relevant cash flow timing, services already provided, group allocations, existing estimates and measurement rules determine the accounting result.
An internal ledger-to-contract-group matrix should have a column for expense eligibility, the governing accounting rule, group identifier, allocation method, reporting-period and amount. An insurer can maintain an attribution file at the individual claim level and aggregate it to groups after verifying contract mapping. For shared eligible overhead, direct claim identifiers may not exist; the group allocation can nevertheless be rational if supported by actual resource usage or another defensible relationship. Applying an arbitrary proportion to groups solely to achieve desired accounting results is incompatible with credible reconciliation.
Actuarial estimate and booked reserve
ASOP No. 43 distinguishes an actuarial unpaid claim estimate and the amount booked as a reserve; the standard covers the adjustment expense component within its stated US professional scope. The 2024 revision of ASOP No. 36 addresses actuarial services for specified written reserve opinions, effective October 1, 2024. These professional standards do not transfer a claims manager's unit-cost calculation into a certified actuarial opinion.
Prospective operational cost projections can support aggregate estimates by identifying cohorts of open claims, work remaining and assumed staffing or service costs. The estimate must also consider relevant claims yet to be reported, reopened claims, reporting and settlement patterns, and other expense components within its scope. A sample of currently open claims omits unreported events and may be enriched with unusually lengthy claims. Extending the sample mean indiscriminately can understate or overstate the remaining liability.
Johnson's (1989) ULAE liabilities study describes in its publisher abstract a method based on reporting and closure patterns that accommodates changing claims-department cost levels. Allen and Mango's (1999) alternative ULAE reserve methods abstract describes alternatives to a simple paid-expense-to-paid-loss ratio and a count-based method. These sources support the conceptual distinction between monetary-loss and workload drivers; their full methods are not reproduced or claimed to have been independently validated here.
An illustrative workload projection consists of 400 known open claims requiring an average of 90 further minutes at $1.50 per minute, or $54,000, and 100 hypothetical unreported claims expected to require 60 minutes each at the same rate, or $9,000. Projected external service cost is $30,000, giving a simplified $93,000 expected resource cost. If the remaining task time on known claims increases by twenty percent, its cost rises to $64,800 and the total becomes $103,800. If the rate instead increases ten percent across both groups, labor becomes $69,300 and the total $99,300. These calculations assume all costs remain applicable and avoid overlaps between groups; they are not a complete actuarial estimate or reserve recommendation.
Recovery and reinsurance
Recoveries and expenses associated with pursuing them require separate transaction categories. A hypothetical $20,000 indemnity payment, $2,000 ordinary adjustment expense, $1,000 recovery expense and $5,000 salvage receipt yields a net case cash outflow of $18,000. The gross cost still includes $3,000 of adjustment and recovery services. A vendor fee must not disappear from cost comparisons merely because the related recovery exceeds it.
Reinsurance reimbursement of claim expenses depends on the treaty and applicable reporting basis. Gross direct cost should remain traceable; the ceded amount follows actual treaty terms, coverage, limits and reporting arrangements. Under the NZ IFRS 17 text, reinsurance-held cash flows are excluded from the direct insurance contract cash flow measurement and accounted for separately. An operational net-cost view is useful for cash management but cannot substitute for separate gross and reinsurance accounting.
Expense Governance and Audit
Accountability and control objectives
Claims expense governance seeks completeness of resources recorded, accuracy of allocation, sound procurement, lawful claim administration and a reliable audit trail. Management controls should address the economic purpose of the expense and the quality of the service. Cost efficiency is evaluated within the obligations governing claims, rather than through a target that competes with the delivery of entitled benefits.
IAIS Insurance Core Principle 19 requires supervisory attention to fair, transparent and timely claims handling; its guidance addresses explanations, competence, conflicts of interest, dispute processes and oversight of outsourced claims work. It is an international supervisory framework, whose domestic implementation and binding effects vary. The UK FCA's ICOBS 8.1.1 R establishes claims-handling requirements within its applicable regulatory scope, including fairness, reasonable assistance, appropriate progress information and timely settlement after agreement. These sources establish why cost indicators require quality and timeliness controls.
A control environment can organize responsibilities around claims operations, vendor management, finance, actuarial functions, compliance and internal audit. Claims staff validate services and attribution to cases. Procurement verifies contracts and commercial terms. Finance controls accruals, expenditure classification, reconciliations and reporting mappings. Actuaries assess the suitability of prospective expense information in relevant estimates. Compliance reviews conduct risk and applicable legal requirements. Independent audit tests control design and operation.
Data structure and transactions
A minimally useful case expense record contains a stable transaction identifier, economic expense category, responsible legal entity, source ledger and account, currency, gross amount, service date or period, booking date, payment date where applicable, claim or other cost-object identifier, vendor or resource pool, evidence identifier, allocation basis version, mapping version and correction history.
Records with missing claim identifiers should remain in a suspense or shared-cost category until resolved. They should not be assigned to a convenient claim solely to complete a report. A distinct aggregate code is appropriate when the service genuinely supports a pool. The suspense balance and age should be reported, with responsibility for remediation.
An invoice should be matched to the authorized engagement and record of service delivery. Appropriate review detects duplicate invoices, reused external reference numbers, credits applied to the wrong vendor, inconsistent currency, invalid tax treatment, and services outside approval authority. Where a vendor submits a lump sum, the organization must determine whether sufficient information exists for claim-level or event-level assignment.
A reproducible allocation run should preserve input snapshot, cost-pool definitions, observed or estimated driver quantities, rate assumptions, versioned calculation logic, exception output and reconciliation totals. If an auditor cannot recreate a period's allocation using the same inputs and rules, the system has only reporting output rather than reproducible financial evidence.
Audit testing and edge cases
A risk-based audit can select samples by amount, vendor, claim type, unusual classification, related-party relationship, reporting cutoff, manual override and suspected duplication. A random component protects against overlooking routine but widespread errors. Audit tests distinguish the existence of the service, accuracy of the amount, appropriate claim assignment, correct timing, classification, authorization and treatment of corrections. Testing only invoice authenticity leaves allocation accuracy unresolved.
Consider three exception cases. A canceled inspection with a $150 contractual cancellation fee consumes expense without a completed inspection. It should be linked to the cancellation or engagement, not dropped from vendor cost. A negative $300 credit for duplicate billing should reverse the originating expense, not be attributed to an unrelated open file. A service shared across five claims with no evidence of relative demand can remain event-level expense; equal assignment is possible as a transparent convention but must not be described as directly measured consumption.
Likewise, a payment returned by the banking system may require a transaction correction and additional reprocessing activity. The returned indemnity payment is not necessarily an adjustment-expense saving; the subsequent manual work is a separate resource cost. A no-payment file with legitimate consultation must still retain its adjustment expense.
Quality-linked metrics
Useful indicators include resource cost per reported original file, expense per closed episode, external fee per validated deliverable, active processing time per task, open-file maintenance expense, cost by comparable complexity band, reopened-file expense, unused capacity, and invoice-accrual aging. All require an explicitly specified period and denominator. For example, expense per original reported file and expense per final closed original file describe different cohorts unless the study period and claim development are matched.
A portfolio might report lower average expense because a larger proportion of straightforward claims arrived in the period. A team might achieve a low average settlement time through an unusually high number of denials. Neither metric independently establishes improved productivity, lawful decisions or service quality.
Quality indicators can include independently sampled decision correctness, unresolved complaint age, notification and payment timeliness, reversal rate, reopenings after correction, explanation completeness and claimant access to assistance. Their design should respect privacy and avoid rewarding exclusion of vulnerable or complex cases. A satisfactory low-cost process completes the legally and contractually necessary work with fewer redundant resources while maintaining the quality of outcomes.
Access and information governance
Claims records may contain medical, legal, financial or other sensitive information. Cost attribution generally requires activity facts and financial identifiers rather than unrestricted access to underlying sensitive content. Analyst views should minimize the data needed for the purpose, use appropriate role separation and preserve controlled linkage to the authoritative evidence. Privileged material should not be copied into a generic operational data warehouse solely to support a cost code.
Data minimization, retention periods, access and cross-border transfer obligations depend on applicable law and contract. This study provides no comprehensive legal determination of these duties. Material changes in data access, vendor location or automated classification should receive appropriate jurisdiction-specific review. Aggregated reports should be checked for reidentification risk when small cohorts or unusual cases are shown.
Comparative Institutional Applications
Property and catastrophe insurance
A standard property claim may require notification, policy verification, inspection, repair valuation and payment. A catastrophe adds event-level coordination, surge staff, temporary premises, transport, specialist mobilization and network management. Operational attribution should distinguish resources consumed by individual claims from readiness or event resources shared by a portfolio. An event cost allocation based on indemnity can concentrate costs in highly damaged properties even when dispatch and field service require similar time for smaller claims.
An illustrative catastrophe contract includes $60,000 for event mobilization and $400 per completed assessment. Five hundred assessments consume $200,000 of transactional service; total event cost is $260,000. Full absorption gives $520 per assessment; marginal contractual service cost is $400 within the assumed capacity arrangement. If only 250 assessments are required, total falls to $160,000 but full cost becomes $640 per assessment. Higher unit cost in the smaller event is a consequence of spreading fixed mobilization over fewer assignments. The study therefore recommends separating readiness and variable usage when evaluating such contracts.
Allocation across coverage and events should retain actual event linkage, catastrophe code version, shared engagement evidence and any service spanning multiple events. Duplicate claims from the same property may be legitimate coverage features or administrative duplicates; the distinction should be checked before calculating cost per event or per claimant.
Motor and liability insurance
Motor claims can combine physical damage, bodily injury, legal liability and recovery. A single event may generate multiple claimant and coverage features with different workflows. Work to determine vehicle damage may be routine while bodily injury evaluation develops over a prolonged period. An event-level average hides the difference; a feature-level model can reveal it but must reconcile common event work separately.
Liability insurance can require defense costs before indemnity is determined, and defense expenditure may persist even after some issues are resolved. Insurance contracts can contain limits, retentions and expense provisions whose interpretation affects the insurer's obligations. The cost ledger should identify the covered benefit, defense of the insured, insurer's coverage analysis and recovery efforts as distinct activities where ascertainable. This research does not determine whether a particular defense expense erodes a policy limit. The original contract and applicable law control.
Vendor budgets for counsel may be set by stages such as initial analysis, discovery, experts, hearings and trial preparation. An estimate should track work completed and expected future tasks. Budget variance may reflect legitimate new evidence or changes in legal strategy; a mechanical refusal to authorize required defense work based solely on variance is not a sound control. Investigation of variance must distinguish unnecessary expenditure from unavoidable complexity.
Health, life and disability insurance
In high-volume health administration, transactional activity can include eligibility checks, coding validation, benefit determination, coordination, exception review and payment. A provider invoice can represent a covered service while an independent review fee represents administration. A mass processing platform may operate with highly skewed exception costs. Cost per claim transaction should not be equated with cost per patient or episode.
Life insurance may have relatively standardized payment processes in some cases and extensive evidence collection in others. Disability claims may involve repeated assessments and duration-dependent maintenance rather than one terminal payment. The unit of work for a continuing disability benefit can be a review episode rather than a closure. Models derived from short-tail property insurance are unlikely to capture that operational structure without adaptation.
ASOP No. 43 explicitly identifies its property/casualty scope and specified health-plan exclusions. Thus its use as a conceptual reference in this dossier does not extend the standard to all health or life claims. The general managerial distinction between activity cost, cash flow and financial eligibility persists across lines; detailed actuarial and regulatory requirements must be established separately.
International comparison
Comparative analysis starts with the institutional role of each document. The NAIC form expresses US statutory presentation categories; it does not establish the globally preferred management-accounting taxonomy. NZ IFRS 17 provides an accounting application of international insurance contract concepts within its jurisdiction; it is not US statutory accounting. UK ICOBS addresses conduct for firms and activities within its scope, not the measurement of an insurance contract group. IAIS provides international supervisory benchmarks, rather than self-executing domestic law.
The general operational problem remains comparable: identify the resources consumed, distinguish the purpose of the work, assign a documented cost object, reconcile the financial amount, and interpret the result within actual contractual and legal requirements. Comparative validity depends on consistent definitions of expense, claim, time, currency, insurance line and stage of development.
A cross-border benchmarking exercise should convert currencies on a documented basis and distinguish exchange-rate effects from labor price differences. Purchasing-power comparisons may answer a different question from nominal financial comparisons. Different legal procedures and service responsibilities can change the mix of activities. Harmonizing the reporting categories without accounting for these underlying differences can produce a standardized table that is economically misleading.
Portfolio Diagnostics and Sensitivity
Mix effects and standardized comparisons
An insurer observes two periods, each with 1,000 claims. In period A, 800 routine files cost $50 each and 200 complex files cost $300 each: $100,000 total, $100 per file. In period B, 900 routine files cost $55 each and 100 complex files cost $330 each: $82,500 total, $82.50 per file. The average expense falls 17.5 percent even though within each type expense rises 10 percent.
At period A's fixed mix, period B's cost becomes 800 times $55 plus 200 times $330, or $110,000, yielding $110 per standardized claim. The apparent portfolio improvement comes from a shift toward routine claims. A decomposition using period A weights attributes a $10 increase to within-category unit cost changes and a $27.50 decrease to the changed mix evaluated at period B rates; the net change is -$17.50. A different decomposition ordering may distribute the interaction differently while preserving the total.
A benchmarking report should therefore show both observed aggregate cost and standardized cost for a common mix, along with the actual volume of each group. Even a well-standardized comparison can be biased if unobserved complexity changes within categories. Confidence intervals or sensitivity bounds can be useful when costs are sampled rather than observed for the full portfolio.
Unit costs, ratios and unstable denominators
An expense-to-indemnity ratio is informative about the relationship between administrative spending and claim payments, but its behavior can mislead when indemnity changes. If $100,000 expense accompanies $1,000,000 indemnity, the ratio is 10 percent. If indemnity doubles and expense is unchanged, the ratio falls to 5 percent even though processing efficiency has not improved. The ratio is undefined when indemnity is zero. For a group containing many denied or investigated no-payment claims, cost per notification and resource usage can remain meaningful.
Ratio averaging also matters. A claim with $1,000 indemnity and $500 expense has 50 percent expense-to-indemnity; another with $9,000 indemnity and $500 expense has about 5.56 percent. The arithmetic mean of individual ratios is approximately 27.78 percent; the portfolio ratio is $1,000 / $10,000 = 10 percent. These statistics answer different questions. Portfolio ratios should use aggregate numerator and denominator, while claim-level distributions should be reported separately with a declared treatment of zero denominators.
Cost per paid claim excludes no-payment investigations if the numerator does too. If the numerator includes all claims-function expense but only paid claims in the denominator, it describes the burden of total processing relative to successful payments. The indicator can be valid for budgeting if named correctly. It is not directly comparable with cost per all notified claims.
Capacity stress and rate changes
Return to the $120,000 labor pool. If eighty thousand practical minutes become seventy thousand because of vacancies and temporary disruptions while the supplied cost remains constant, the rate becomes about $1.71429 per minute. At unchanged consumption of 65,000 minutes, assigned labor cost becomes about $111,428.57 and residual capacity cost $8,571.43. Under the original rate, assigned labor was $97,500 and residual $22,500. The differences reflect changed capacity. Operational productivity should be examined separately through tasks completed per practical hour and quality.
If demand reaches 90,000 minutes when practical capacity is 80,000, the original rate produces $135,000 assigned cost against $120,000 supplied. Negative unused capacity is a diagnostic of an inconsistent assumption, overtime or excess estimated workload; it is not a literal negative expense. Management must reconcile actual overtime, contracted resources, changes to standards, workload measurement and deferred activity before interpreting the result.
For forecasting, scenarios can vary claim arrival, shares of complex files, third-party prices, task durations and available capacity jointly. If vendor prices rise during a catastrophe at the same time as inspections become longer, treating those drivers as independent may understate cost volatility. Stress assumptions should be explicit. Bounds derived from historical data require actual observed distributions; synthetic shocks are useful only as conditional illustrations.
Forecast monitoring and validation
A prospective estimate can be tested against subsequent service payments and actual resource usage. The comparison should hold the original forecast date and information set fixed. Separating price, volume, mix, task time, capacity and timing variance helps identify whether errors arise from economic assumptions or ledger settlement delays.
For example, suppose a forecast anticipates 200 inspections at $300 each, or $60,000. Actual volume is 220 at $330, giving $72,600. Using forecast price first, volume variance is 20 times $300 = $6,000. Price variance evaluated at actual volume is 220 times $30 = $6,600. The two sum to $12,600. When services differ in scope, however, part of the price variance may represent complexity or scope mix. The analytical decomposition should not be presented as causal proof.
Monitoring should include cohorts that have not fully developed. Closed-file-only cost samples can exclude long-running expensive cases, while recent reported cohorts may have incomplete expense development. A survival-aware or development-based analysis can estimate the progression of costs over time when suitable data exist; its assumptions require independent validation. This dossier demonstrates the importance of that distinction but does not fit a survival model.
Integrated Analytical Synthesis
The research establishes four related analytical conclusions within existing cost-accounting and claims disciplines. First, expense attribution and financial recognition require independent classifications that can be linked through explicit, reviewable mappings. Traceability to a case is a managerial data property; eligibility for financial measurement is an accounting assessment under the applicable framework. Combining them into one code prevents either distinction from being audited.
Second, the economic interpretation of a cost depends on the resource boundary and decision horizon. Full attribution measures the share of supplied resources consumed or absorbed, while incremental expenditure describes a change in organizational cash or obligations. The controlled capacity example shows that an identical complex claim can receive different assigned costs solely because a denominator changes. The vendor retainer example shows that a per-inspection margin and a fully absorbed service cost can both be accurate and support different decisions.
Third, resolution pathway cost is conditional on both case composition and the sequence of activities. Litigation-related cost and duration cannot be inferred reliably from raw observed differences among groups whose underlying cases differ. The synthetic expected-value exercise establishes a sensitivity relationship between probabilities and prospective outflows. It also illustrates why sunk costs and the insurer's net economic exposure must not be confused with contractual entitlement. These propositions follow from standard decision analysis and are not proposed as a new theoretical model.
Fourth, the validity of operational incentives rests on the service and outcome boundary. Lower administrative expense would be an incomplete objective if obtained through incomplete investigation, inaccessible processes or unjustified delay. Timeliness, decision correctness, dispute handling, complaint outcomes and oversight of external services should be evaluated alongside the ledger. This is an interpretation of cost measurement under existing claims governance requirements, rather than a new doctrine of claims conduct.
Taken together, these conclusions identify a practical architecture for research and institutional analysis: preserve original financial transactions and service evidence, classify by economic purpose, attribute according to documented resource relationships, isolate residual capacity, test comparative assumptions, reconcile accounting views and monitor the resulting incentives. Every aggregation must retain its path back to original definitions and records. Precision in a cost model cannot restore context that was discarded during data collection.
Research Limitations
This dossier is a documentary and analytical study. Its numerical examples are constructed, reproducible arithmetic rather than sampled insurer experience. Empirical claims are limited to the content of the accessible publication summaries or abstracts. No universal parameter, causal effect, market-average cost or representative insurer behavior is inferred. Source selection is purposive rather than a comprehensive systematic review of every language and jurisdiction.
The complete texts of the Johnson and Allen-Mango actuarial papers and Asmat-Tennyson working paper were not reviewed. Their abstracts are used only for the specific propositions stated there. The Insurance Research Council report summary does not provide independent access to underlying claims data. The 2026 NAIC instructions were consulted through extracted text; a requested page image was unavailable. The 2025 form cover and extracted expense rows were examined. The detailed IFRS 17 cash flow discussion relies on official NZ IFRS 17 text and the IFRS Foundation educational definitions, rather than subscription-based full IASB text. The chapter avoids a binding filing conclusion for an unidentified entity or state.
No original case files, policy wording, vendor contracts, invoices, claimant records or insurer ledgers were examined. Worked cases accordingly do not settle actual claims, certify reserves, recommend a particular indemnity or establish compliance by a real firm. Empirical implementation and a full multi-jurisdictional legal survey remain outside the achieved scope. These boundaries are central to the interpretation of the work and are not presented as new findings.
Conclusion
Claims expense attribution explains the costs of administering and resolving insurance obligations by relating actual and estimated resource consumption to identifiable activities and cost objects. Its scientific value depends on transparent distinctions among services, legal entitlements, transactions, resource capacity, time periods and financial measurements.
Activity-based costing and time-driven allocation can provide defensible representations of service intensity when rates, driver quantities and capacity are documented and validated. Shared costs must remain visible when a causal allocation is unavailable. Direct vendor linkage improves traceability but does not settle questions of expense purpose, incremental cost, or accounting treatment. The same principle applies to defense activities, technology and recovery services.
Settlement pathway models can organize the uncertain financial consequences of feasible options; their assumptions must remain explicit and be tested against changes in volume, composition, probabilities and costs. Claimant rights, fair treatment and competent assessment establish constraints on the operation of those pathways. Ledger reconciliation preserves total expenditures and outstanding balances while maintaining a separate documented interface to actuarial and financial reporting requirements.
The resulting analysis positions settlement cost accounting as an indispensable component of claims science. Well-designed attribution reveals the resource demands and operational tradeoffs underlying a claims system, while faithful reconciliation and governance make those results interpretable and accountable. The object of cost improvement is the elimination of redundant resource consumption consistent with the correct and timely fulfillment of insurance obligations.
References
Actuarial Standards Board. (2007). Property/Casualty Unpaid Claim Estimates. Actuarial Standard of Practice No. 43. Adopted June 2007; effective September 1, 2007. Document.
Actuarial Standards Board. (2024). Statements of Actuarial Opinion Regarding Property/Casualty Loss, Loss Adjustment Expense, or Other Reserves. ASOP No. 36. March 2024; effective October 1, 2024. Document.
Allen, C. A., & Mango, D. F. (1999). Two Alternative Methods for Calculating the Unallocated Loss Adjustment Expense Reserve. Casualty Actuarial Society E-Forum, Fall, 225-250. Document.
Asmat, D., & Tennyson, S. (2016, January). Tort Liability and Settlement Failure: Evidence on Litigated Auto Insurance Claims. Economic Analysis Group Discussion Paper EAG 16-1. U.S. Department of Justice, Antitrust Division. Document.
External Reporting Board. (n.d.). NZ IFRS 17 Insurance Contracts. Standards Navigator, consolidated online text, paragraphs B61-B66. Document.
Financial Conduct Authority. (n.d.). Insurance: Conduct of Business Sourcebook (ICOBS), Chapter 8, Claims Handling, rule 8.1.1 R. Document.
HM Revenue & Customs. (n.d.). General Insurance Manual: Glossary (GIM99999). GOV.UK. Document.
IFRS Foundation. (n.d.). IFRS 17 Insurance Contracts: Key Terms. Educational implementation material. Document.
Insurance Research Council. (2026, July 30). Auto Injury Insurance Claims: A Study of Increasing Claim Severity. Research Report Summary. Document.
International Association of Insurance Supervisors. (n.d.). Insurance Core Principles and Common Framework for the Supervision of Internationally Active Insurance Groups. ICP and ComFrame Online Tool, ICP 19.10-19.11. Document.
Johnson, W. A. (1989). Determination of Outstanding Liabilities for Unallocated Loss Adjustment Expenses. Proceedings of the Casualty Actuarial Society, 76, 111-125. Document.
Kaplan, R. S. (2005, January 24). Rethinking Activity-Based Costing. Harvard Business School Working Knowledge. Excerpt of Kaplan and Anderson (2004), Time-Driven Activity-Based Costing, Harvard Business Review, 82(11), 131-138, 150. Document.
National Association of Insurance Commissioners. (2025). Annual Statement Blank: Property/Casualty. 2025 edition. Document.
National Association of Insurance Commissioners. (2026). Property/Casualty Annual Statement Instructions. Underwriting and Investment Exhibit, Part 3, p. 94. Document.