Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Health and Disability Reinsurance Programs

Abstract. Health and disability reinsurance programs allocate the financial consequences of medical expenditure, prolonged incapacity, income-protection claims, and correlated morbidity between primary insurers and specialized risk-bearing counterparties. Their effectiveness depends on the relation between underlying benefit obligations, the statistical distribution of claim amounts and durations, contract-specific reinsurance triggers, and the enforceability and collectibility of recoveries. This monograph investigates those relations across medical expense insurance, short- and long-term disability insurance, income protection, and adjacent arrangements. It derives the mechanics of proportional, per-person excess-of-loss, aggregate stop-loss, and duration-sensitive reinsurance through explicit mathematical illustrations; examines how high-cost claimants, recovery rates, health-system price inflation, provider networks, infectious disease events, and shared operational exposures affect both expected losses and tail outcomes; and compares relevant international supervisory, accounting, and actuarial frameworks. The source base combines the International Association of Insurance Supervisors' Insurance Core Principles, the European Union's Solvency II framework, IFRS 17 materials, Canadian reinsurance guidance, U.S. insurance and health-benefit regulation, actuarial standards, peer-reviewed health-economics literature, and published experience studies of medical costs and disability. The examination of 2022 U.S. medical expenditure concentration and the 2015–2022 and 2015–2024 group long-term disability experience studies is expressly limited to their populations, measurement conventions, and publication dates. Publicly documented U.S. health risk adjustment and state reinsurance arrangements are analyzed as institutional comparators; their payment mechanisms and statutory beneficiaries differ from ordinary bilateral insurance treaties. The central finding is that medically defined severity, contractual ceded severity, and realizable recovery must be measured separately. A reinsurance design can materially reduce an insurer's nominal tail loss while leaving duration risk, claim-definition mismatch, calendar-year leakage, counterparty default, provider concentration, or liquidity stress. Program appraisal therefore requires simultaneous examination of benefit definitions, exposed populations, stochastic morbidity, claim development, treaty wording, settlement timing, capital recognition, and governance. The research contributes a source-grounded comparative synthesis and transparent actuarial derivations; it makes no claim to a newly established theory or empirical causal discovery.

Keywords: Health reinsurance; disability reinsurance; medical expense insurance; long-term disability; morbidity; claim duration; specific excess of loss; aggregate stop-loss; quota share; high-cost claimants; health catastrophe; provider concentration; risk adjustment; reinsurance recoverables; IFRS 17; Solvency II; actuarial risk transfer; counterparty risk.

Introduction

A health or disability insurer promises financial performance under a primary contract whose ultimate cost depends on events beyond the time of underwriting. Medical coverage may finance a sequence of hospital admissions, pharmaceuticals, specialist interventions, or continuing treatment; disability coverage may promise income replacement for months or decades after an insured becomes unable to work. Reinsurance provides a second contractual relationship through which part of those contingent costs is indemnified or otherwise transferred. The legal obligation to the original policyholder ordinarily remains with the direct insurer. The distinction is foundational in the IAIS treatment of reinsurance and in the IFRS guidance on reinsurance held.

Health and disability risks have a complicated exposure architecture. A high-cost claimant can generate very large benefits within one plan year and substantial subsequent costs in later years; an initially modest disability claim can become expensive through persistent benefit payments; a pandemic can affect numerous policyholders, medical providers, employees, and claim-administration systems simultaneously. The direct insurer's capacity to respond reflects its reserves, available capital, provider-payment terms, liquidity, and reinsurance. A contractual layer that pays only after a specified threshold can absorb one form of volatility yet leave other forms untouched. Consequently, the evaluation unit must be specified before an estimated recovery or capital benefit can be interpreted.

The economic study of medical insurance begins from uncertainty about illness and the effectiveness of treatment. Arrow's (1963) analysis of medical-care uncertainty connects these contingencies to the institutions of health financing. Cutler and Zeckhauser's (2000) account of health-insurance design demonstrates why risk sharing interacts with incentives, medical-service decisions, and selection. Reinsurance operates at a different contracting level but inherits the underlying benefit design: it covers financial obligations arising from services and benefits that the primary insurance contract has already defined.

This research concerns programs insuring or reinsuring medical expense and disability obligations. It includes proportional treaties, specific excess protection, aggregate covers, and carefully delimited comparisons with employer stop-loss insurance, public risk adjustment, and health-event financing. Its principal Science is Reinsurance; its related Science are Health Insurance, Long-Term Care and Disability Insurance, and Actuarial Science. The approved research identity is the H1 above and the single /dossiers/ canonical route. The investigation is international in scientific coverage, while statutory statements retain their named jurisdictions and effective versions.

The analytical method moves from the unit of insured exposure to gross loss distribution, contract transformation of loss, timing of recoveries, and financial recognition. Each numerical example is hypothetical and deliberately reproducible. Source-grounded empirical facts are distinguished from mathematical identities and economic interpretation. The final comparison establishes when risk transfer meaningfully alters retained losses, when nominal capacity overstates effective protection, and how contract drafting and data governance determine the outcome.

Scientific Definitions, Scope, and Sources

Reinsurance is a contract under which an insurer, the cedent, obtains protection from a reinsurer against agreed exposures arising from the cedent's insurance obligations. The reinsurer accepts specified risk in exchange for consideration, subject to definitions, attachment, exclusions, duration, limits, and settlement provisions. The NAIC's reinsurance overview describes the relationship and the existence of retrocession, through which a reinsurer itself acquires protection. ICP 13 treats reinsurance as a means of underwriting-risk transfer that simultaneously creates credit, operational, and basis risk. The fact that two firms agree on a ceded percentage or a recovery limit does not by itself establish the timing or full economic value of recovery.

Health insurance here means financial coverage for the costs of defined medical services or benefits. Disability insurance includes benefits contingent on sickness, injury, impaired capacity for work, or a contractually defined disability status. Income protection replaces some earnings when the insured satisfies a disability definition and any elimination period. Long-term care primarily addresses assistance needs and care dependencies; overlapping morbidity methods do not make long-term care and disability income contracts identical. Health and disability coverage can be issued by life insurers, health insurers, or other authorized entities, depending on the jurisdiction and contract. The corresponding reinsurance may fall within different regulatory lines and capital calculations.

Four related measurements must be maintained. Gross insured loss is the direct insurer's covered obligation before reinsurance; ceded insured loss is the part eligible under the reinsurance contract; contractual recovery is the amount calculated from eligible ceded loss after retentions, exclusions, limits, and reinstatement provisions; collectible recovery is the amount reasonably expected to be paid by the reinsurer, considering counterparty and settlement risk. A fifth measurement, cash recovery by time, determines whether the insurer can fund actual primary claim payments. These distinctions are analytical definitions applied throughout the dossier. They do not constitute a new theoretical system.

The unit of risk varies by treaty. It can be an individual covered person, an episode, a disability claim, a policy, an employer group, a risk pool, an occurrence, or an accounting period. A person-specific treaty may count a claimant's annual eligible medical expenditures while another counts all paid claims arising from one condition; a group disability treaty may allocate liability according to date of disability onset, policy issuance, or a specified accounting cohort. Those contract rules are substantive, not merely administrative. A reinsurance recovery is meaningful only when the underlying event, coverage dates, beneficiary, and eligible amounts can be reconstructed from records.

The economic and legal identity of stop-loss coverage requires care. An insurer reinsuring a licensed insurer assumes a different contract from an insurer covering a self-funded employer's liability under a health plan. The NAIC's (2015) stop-loss and self-funding white paper expressly explains this distinction. The U.S. Department of Labor's (2014) guidance on stop-loss regulation addresses the interaction between self-funded ERISA plans and state regulation of insurance policies sold to their sponsors. Although identical mathematical loss functions may appear in a specific stop-loss policy and a health reinsurance treaty, their buyers, insured obligations, legal classification, credit treatment, and consumer-protection context can differ.

Methodologically, international regulations are treated as jurisdiction-specific law or supervisory standards, not as interchangeable global commands. The IAIS establishes a globally used supervisory framework; its Insurance Core Principles are implemented through jurisdictional systems. IFRS 17 governs reporting by entities within its scope rather than determining each country's solvency capital by itself. Experience studies are samples with defined inclusion rules, not universal disease or employment statistics. The present work therefore derives general actuarial identities while limiting historical, empirical, and regulatory assertions to the evidence that supports them.

The Risk Structure of Medical and Disability Obligations

Medical costs combine the frequency of covered services, their severity, the intensity of treatment, and the prices paid to providers. Let N denote a count of covered episodes and X_i the allowed cost of episode i. In a simplified compound-loss representation, the total S equals the sum of X_i for i from 1 to N. Actual health claim systems must also account for the contract's deductibles, copayments, coinsurance, provider adjustments, coordination of benefits, exclusions, and policy limits. A reinsurance loss record based on gross billed charges will therefore differ from one based on the primary insurer's final eligible payments. Any stochastic representation must declare which of these quantities is modeled.

The upper tail of medical expenditures deserves particular attention. Hernandez-Viver and Mitchell's (2025) AHRQ expenditure concentration study estimates that the highest-spending one percent of people accounted for 21.7% of U.S. civilian noninstitutionalized medical expenditures in 2022, and the highest-spending five percent accounted for 49.7%. The study measures all-payer personal health expenditures in its specific survey population. It does not report a reinsurer's eligible paid claims, policy-specific loss ratio, or international tail factor. Transferring those population statistics directly into a treaty price would confuse expenditure concentration with ceded-claim concentration. Nevertheless, the observation justifies testing whether a relatively small set of insured persons can dominate the upper layers of a medical portfolio.

The risk of a large medical claimant has several temporal dimensions. A costly specialty drug may produce repeated expenditures; a transplant can involve a concentrated procedure followed by continuing care; a complicated neonatal case can generate substantial expenses at several providers; a chronic disorder may yield multiple high-cost periods rather than one claim. A per-person annual cover responds to the sum of treaty-eligible expenses within its annual window, whereas per-event cover depends on an episode definition. Cost inflation can push otherwise unchanged claimants across a fixed attachment point. Changes in provider contracting can alter paid severity without an equivalent change in morbidity.

Disability has a distinct incidence-duration structure. At onset, the insurer assesses whether the definition of disability is met, whether the waiting period is satisfied, and whether the claimant is eligible for a recurring benefit. Ultimate liability depends on recovery, death, retirement, return to work, offsets with social-insurance benefits, partial-disability definitions, and contract termination rules. A pool with stable new-claim incidence can still experience substantially higher ultimate cost if disability durations lengthen. In addition, two insurers with similar current-year cash payments may have very different future benefit commitments because their open-claim duration distributions differ.

The Society of Actuaries Research Institute and LIMRA (2025) describe 2015–2022 group disability incidence based on about 294 million life-years and 1.2 million observed claims from 19 contributing companies. Their report restricts the study to specified fully insured group LTD business and excludes, among other categories, administrative-services-only business and certain short-duration contracts. The SOA Research Institute and LIMRA's (2026) 2015–2024 termination experience reflects a different study design: more than 43 million life-months exposed, approximately 937,000 recoveries, and more than 181,000 deaths reported by 18 contributors. The first study informs entry into disability; the second informs departure from the disabled state. Their exposure denominators, contributing insurers, years, and eligibility screens must be reconciled before combining estimated rates.

A health-disability reinsurance program that covers both medical and income benefits faces correlation beyond any single claim type. Serious illness can simultaneously increase treatment costs and remove workers from employment; healthcare-provider interruption can delay treatment and recovery; administrative backlogs can extend claims even after physical impairment changes. Such relationships should be represented through joint scenarios and conditional transitions where reliable data allow, and through sensitivity analysis where data do not. The health-economics literature also warns that utilization and benefit design respond to incentives: the Cutler–Zeckhauser analysis addresses the interplay of risk pooling, adverse selection, moral hazard, and medical decisions.

Insurance distribution also affects which claimants reach a reinsurer. Differences in age, occupation, geographic location, benefit maximums, provider contracts, diagnoses, and underwriting practices shape the reinsured portfolio. Finkelstein and McGarry's (2006) study of private information in long-term care insurance demonstrates that observed insurance demand and underlying private risk may reflect multiple dimensions of information. The study's empirical identification is specific to its market and data. It supports attention to selection and behavioral heterogeneity, but supplies no direct coefficient for health or disability reinsurance pricing.

High-Cost Claim Protection and Treaty Attachment

Specific excess-of-loss reinsurance protects against eligible loss attributed to a defined individual risk or claimant above a retention. Let C_j be the treaty-eligible accumulated claim amount for claimant j during the applicable contract period, A the per-claimant attachment, and L the maximum recovery for that claimant. The purely financial recovery function is:

Rj=min⁡{max⁡(Cj−A,0),L}.R_j=\min\{\max(C_j-A,0),L\}.

This function must be applied only after the contract's eligibility, timing, and exclusion rules. If A is $250,000, L is $750,000, and a claimant produces $1,000,000 of eligible claims during the specified period, the calculated recovery is $750,000 and the cedent retains $250,000. If the eligible claim is $1,400,000, recovery remains $750,000 under this capped example and the cedent retains $650,000. Both results follow from the stated function; neither reflects actual market treaty terms. A policy with an unlimited per-person maximum or a separately specified annual aggregate cap gives a different result.

The distribution of recovery depends on the upper tail of C_j rather than simply its mean. An attachment just above the mean produces many small or moderate recoveries; a high attachment produces infrequent, potentially large payments. If a loss distribution has uncertain extreme values, estimates of E[R_j] can be unstable because few observations fall beyond the attachment. Credibility adjustments, exposure stratification, outlier examination, and explicit tail assumptions become essential. The ASOP 23 data-quality guidance is relevant to the quality, limitations, and use of claims and eligibility data; ASOP 56 on modeling provides a professional framework for model selection, assumptions, testing, and communication within its jurisdictional scope.

Aggregate stop-loss arrangements protect total eligible portfolio loss in a defined period. Let S denote aggregated treaty-eligible losses, B the aggregate attachment, and M the aggregate cover limit. A simple recovery is:

Ragg=min⁡{max⁡(S−B,0),M}.R_{\mathrm{agg}}=\min\{\max(S-B,0),M\}.

Suppose a reinsured health portfolio incurs $8 million of eligible annual loss, with a $7 million attachment and a $2 million limit. The aggregate cover pays $1 million under the stated assumptions. If total loss were $10 million, the same cover would pay $2 million and leave $1 million above the exhausted layer, in addition to the retained amount below attachment. An aggregate treaty can cover the collective effect of medium-sized claims that never reach a high specific attachment. It can also absorb systemic medical inflation or unexpectedly high episode counts, subject to any contract exclusions and sublimits.

Specific and aggregate reinsurance may coexist. Their order of application determines whether the same underlying expense counts twice. If the aggregate base is calculated net of specific recoveries, the net aggregate attachment and the specific treaty's contribution to the portfolio must be coherently defined; if the aggregate treaty uses gross loss but credits specific recoveries differently, the cash result changes. The contract should identify the sequence, recoverable amounts, exhaustion treatment, and any common limits. A modeled program that simply adds hypothetical specific and aggregate recoveries without checking priority can materially overstate protection.

Proportional quota-share reinsurance transfers a stated share q of eligible premium and losses, usually together with defined commissions or expense arrangements. Ignoring expenses, commissions, and exclusions for the illustration, gross eligible loss S of $8 million under a 25% quota share yields $2 million of ceded claims and $6 million of retained claims. The proportional share does not require the loss to exceed a per-person attachment. Its economic cost, commission terms, and capital effect are distinct from those of nonproportional coverage; it can also alter underwriting incentives. Surplus-share and facultative structures may be relevant to certain specialty or large-limit benefits, but their legal bases and treaty mechanics should not be assumed to match property insurance without contract evidence.

Medical expense claims develop across multiple accounting windows. A person may receive treatment in December, the provider may submit the claim in January, the insurer may adjudicate it in March, and the reinsurer may settle later. A treaty attaching on incurred, paid, reported, or another defined basis can allocate this claim differently. Run-in and run-out clauses, prior-year adjustments, annual resets, and changes in enrollment or beneficiary identification determine whether the recovery assumed in a model is actually covered. These are particularly important in high-cost recurring treatment and disability claims that span contract renewals.

The attachment must also be expressed in the correct monetary base. A $250,000 threshold measured in provider charges is economically different from a threshold measured in allowed charges or net covered payments. Currency fluctuations affect cross-border covers; exchange rates should be specified at attachment, valuation, and settlement. Medical benefit caps and reinsurance limits must be consistent in inflation assumptions. The methodologically sound comparison estimates the treaty payout under a common, clearly identified loss basis, then evaluates net retained loss and the timing of recovery under each proposed program structure.

A public high-cost risk pool can resemble excess insurance mathematically without being an ordinary reinsurance treaty. The U.S. Centers for Medicare & Medicaid Services' high-cost risk-pool audit description identifies HCRP as part of HHS-operated risk adjustment from the 2018 benefit year, covering specified individual and small-group markets and requiring validated enrollee-level information. The governmental allocation formula, participants, funding mechanism, and statutory basis must be examined independently. Its existence demonstrates that high-cost claimant sharing can be institutionalized outside the bilateral treaty form.

Disability Morbidity, Recovery, and Duration

Disability risk has two broad actuarial components: the rate at which insured people become claimants and the rate at which existing claimants leave benefit-paying status. An incidence model identifies the hazard of a newly qualifying disability after taking account of the covered population, benefit definition, age, occupation, elimination period, and observation period. A termination model estimates recovery, death, expiration of benefit term, or other exit states conditional on claim duration and claimant characteristics. Reinsurance programs can take a share of both components, a share of incurred claims, or a defined part of outstanding disabled-life liability.

A simple multistate representation has an active state A, a disabled state D, a recovered or active-again state R, and a dead state M. Transitions may include A to D, D to R, D to M, and, where the product requires it, recurrence from R to D. In disability income products, partial disability, residual benefits, unemployment, retirement, and changing occupation may require additional states. The mathematical literature of Haberman and Pitacco (1998), Actuarial Models for Disability Insurance, develops multiple-state approaches for disability and related personal insurance. These methods clarify why a single mean claim duration cannot represent all contingent obligations.

For a hypothetical claim already in disabled status, suppose constant monthly recovery intensity ρ\rho and death intensity μ\mu, with independent competing exits within this stylized continuous-time illustration. The probability that benefits continue to month t, before any other policy termination, is:

P(Dt∣D0)=e−(ρ+μ)t.P(D_t\mid D_0)=e^{-(\rho+\mu)t}.

The expected number of paid months within a horizon H, ignoring elimination periods, discounting, re-entry, and contractual offsets, is:

E[TH]=∫0He−(ρ+μ)t dt=1−e−(ρ+μ)Hρ+μ.E[T_H]=\int_0^H e^{-(\rho+\mu)t}\,dt =\frac{1-e^{-(\rho+\mu)H}}{\rho+\mu}.

For ρ=0.04\rho=0.04 and μ=0.002\mu=0.002 monthly with H = 12, the result is approximately 9.43 expected months. If recovery intensity declines to 0.02 while death intensity remains 0.002, the expected months increase to approximately 10.55. This calculation is original explanatory arithmetic, not a fitted probability of termination from an actual insurer's portfolio. It demonstrates a fundamental sensitivity: duration deterioration can raise expected benefit outlay without any increase in disability incidence. Real valuation adds benefit amount, indexation, discounting, policy limitations, offset rules, and claim-dependent transition patterns.

The SOA 2012 group LTD valuation materials document a valuation framework developed from earlier group experience. The later SOA–LIMRA incidence study and termination study show why assumptions should be continuously compared with emerging experience rather than imported mechanically from a single table. Their industry experience measures claims under defined U.S. group products; an international portfolio must make its own allowance for social insurance, disability adjudication, medical rehabilitation, occupational composition, and benefit design.

Disability reinsurance requires clarity about whether the reinsurer pays its share when disability first attaches, as individual benefit installments emerge, or when a defined reserve amount is transferred. Under a quota-share structure the reinsurer may follow a proportion of covered installments, subject to treaty terms. Under a per-life or per-claim excess arrangement, large present values or unusual durations can be the protected tail. Under a financial reinsurance or block transaction, the nature and extent of genuinely transferred insurance risk may be more complex. The ASOP 11 treatment of health-related risk transfer instructs actuaries performing relevant financial-reporting services to consider the characteristics of these arrangements; it is a professional practice standard, not an assertion that every arrangement earns identical accounting or capital treatment.

Disability claims are also sensitive to service provision and administration. Delayed medical assessment can prolong adjudication; rehabilitation access and return-to-work services may change recovery patterns; contractual offsets related to workers' compensation, social security, or other income sources can change gross benefit obligations. Claim classification by mental, musculoskeletal, neurological, or other diagnoses must reflect the actual product and applicable law. The method must distinguish a genuine change in expected future benefit payments from a change in reporting speed or case management. A reinsurer receiving only aggregate cash values cannot reliably analyze that distinction.

When claim duration extends for years, reinsurance creates a long chain of performance exposure. A cedent can pay benefits monthly while recovering quarterly or after periodic bordereaux. The treaty may terminate, commute, or renew while open disabilities continue. Retention of in-force claims under a claims-incurred basis, the financing of incurred-but-not-reported disabilities, and the rights of both parties upon change of control or insolvency must be established before the expected recovery has operational meaning. Duration-sensitive reinsurance can reduce volatility, but its execution depends on medical, actuarial, administrative, and legal data being consistent over long periods.

Portfolio, Provider, and Health-Catastrophe Accumulation

Traditional portfolio diversification relies on the extent to which individual losses move independently. For losses X_1 through X_n, the variance of aggregate S is:

Var⁡(S)=∑iVar⁡(Xi)+2∑i<jCov⁡(Xi,Xj).\operatorname{Var}(S)=\sum_i\operatorname{Var}(X_i) +2\sum_{i<j}\operatorname{Cov}(X_i,X_j).

The covariance terms can dominate diversification benefits when a portfolio shares geographic, occupational, provider, biological, or policy-design exposures. Adding independent insured people can reduce the relative fluctuation of ordinary claims; adding people exposed to the same outbreak, provider system, or reimbursement change may preserve a substantial common loss component. Health reinsurance therefore requires both ordinary claim simulation and accumulation analysis.

A provider network is a potential source of concentration even when policyholders live in different locations. One large hospital system may treat a disproportionate share of complex cases; a single pharmaceutical class may account for a large share of specialty-drug spending; one claims processor may serve multiple insurers; payment disputes or cyber outages can delay numerous reimbursements simultaneously. A model that groups exposures only by insured person can overlook common providers, disease pathways, benefit managers, employer sectors, or administrative platforms. A useful concentration report shows the largest correlated groups, conditional loss distributions, and exposures relative to reinsurance limits.

Medical catastrophes also require a precise definition of the insured event. A disease outbreak can generate many separate medical claims over months. A treaty based on one occurrence needs an operative event definition, aggregation method, territorial scope, time window, and exclusions. Disability claims arising from a health event may begin weeks after the initial illness and can continue years beyond the event. An annual aggregate health cover may respond more readily to such a spread-out pattern than an occurrence-based treaty with a narrow hours clause, although contractual terms ultimately control. Event severity and payment duration therefore should be modeled on separate timelines.

An accumulation analysis can use scenarios: an outbreak increasing short-term hospitalization and long-term work incapacity; a major hospital-network disruption displacing treatment and altering prices; a supply shock driving critical-drug spending; or a geographically concentrated employer portfolio facing simultaneous disability claims. For each scenario, the analyst should estimate gross insured obligations, eligible ceded amounts, attachment exhaustion, treaty exclusions, reinstatement availability, and expected recovery timing. The dependence assumptions, calibration dates, missing exposures, and sensitivity ranges need explicit disclosure. Under ASOP 56 modeling guidance, suitability and limitations of models are professional considerations for actuaries within the standard's scope.

Public pandemic financing offers a boundary case. The World Bank's Pandemic Emergency Financing Facility account reports a parametric insurance window involving catastrophe bonds and swaps, an allocation of $195.84 million for COVID-19 response, and the facility's closure on April 30, 2021. The facility transferred outbreak-trigger risk for public response financing, rather than reinsuring each primary health insurer's medical or disability claims. It illustrates the possibility of risk transfer using disease-event metrics while also exposing basis risk: a defined payout trigger can diverge from a particular payer's actual medical or disability losses. The recorded experience should not be treated as an estimate of private health reinsurance claims or as a current operating program.

Layer exhaustion creates different vulnerabilities under concentrated scenarios. A per-person cover may exhaust on several extreme claimants while leaving many medium-sized claims unprotected. An aggregate stop-loss cover can run out after an early wave of claims, making later waves wholly retained. A quota share reduces every eligible claim proportionately but places a broad share of underwriting experience with the reinsurer. A financially sound composite can combine these features if its treaties specify interaction and recovery priority. Reinstatement premiums and aggregate annual limits should appear in both the pricing and stress-test cash flows.

Program Architecture, Wording, and Claims Administration

Effective program architecture begins with the insurance portfolio rather than the available reinsurance quotation. It identifies primary policy obligations, the desired retained volatility, statutory and economic capital needs, liquidity limits, and operational capacity. The cedent then evaluates candidate structures against the same exposure dataset and risk scenarios. A treaty should be assessed as a transformation of the underlying loss distribution, not simply as a nominal limit purchased from a rated counterparty.

There are distinct reasons to use proportional and nonproportional treaties. Proportional protection may support growth, sharing of technical risk, or protection across an entire risk class, but it can also transfer profitable ordinary experience and change underwriting incentives. Specific excess protection targets individual outliers and may better preserve ordinary underwriting participation. Aggregate coverage targets portfolio-level adverse development caused by incidence, severity, inflation, or accumulation. The choice requires examining the entire distribution of retained outcomes, the cost of ceded premium, prospective claims-adjustment expenses, credit exposure, and solvency recognition. Ranking alternatives on gross ceded premium alone yields a deficient comparison.

Treaty wording determines what losses the financial structure actually includes. Definitions should identify the reinsured policies, attached lives, eligible benefit classes, reporting units, claim occurrence and incidence dates, attachment basis, allowable medical payments, adjustments, sublimits, exclusions, aggregate caps, commutations, and currency conventions. Disability treaties require the meaning of total and partial disability, offsets, elimination periods, benefit duration, and treatment of reopened or recurrent claims. Health treaties require coordination of benefits, network arrangements, claim recoveries, subrogation, pharmacy rebates, adjustments, and disputes over allowed amounts. The contract should also specify whether underwriting changes, medical inflation, and changes in law alter eligibility.

Notice, claims cooperation, and control provisions must be reconciled with practical claim administration. A cedent generally remains responsible for claims owed under its primary policies and must have the funds and authority to perform, even when the reinsurer disputes reimbursement. IFRS Foundation educational materials on separate reinsurance contract measurement emphasize the distinct performance risks of underlying and ceded contracts. Contractual cash calls, accelerated settlements, claims funds, collateral, or other agreed arrangements may mitigate cash-flow lags; their actual efficacy depends on enforceability and funding.

Bordereaux are a major control point. A health reinsurance bordereau may require anonymized person identifiers, coverage dates, service dates, incurred/paid amounts, benefit categories, provider grouping, historical adjustments, and a trail from gross to eligible ceded amounts. A disability bordereau may require claim onset, decision date, elimination period, payment amounts, reserve movements, termination or recovery state, diagnosis category, reopening rules, and deductions. Data-minimization and privacy rules apply to health information: an actuary requires robust linkage and auditability, while contract participants must restrict personal data to permitted purposes. The scientific point is the need for accurate and privacy-respecting claim identification, not a prescription for any jurisdiction's personal-data processing.

Reconciliation should operate in several directions. First, policy administration must tie to insurance eligibility and premium. Second, claim events must tie to actual ledger payments and outstanding liability. Third, the reinsurance calculation must tie to treaty thresholds and exclusions. Fourth, expected and received recoveries must be reconciled to counterparty statements and financial accounting. Five, bordereaux revisions must retain their provenance so prior submissions can be reconstructed. These controls matter disproportionately where a small number of unusually expensive claims determine a large part of the treaty value.

The IAIS ICP 13 supervisory framework expects insurers to manage their reinsurance programs and associated risks. Canada's OSFI (2025) sets out sound reinsurance practice expectations, including documented strategy, ceding limits, concentration controls, reinsurer due diligence, clear written contractual terms, and monitoring. These are specific supervisory sources with different institutional status, but both reinforce the operational conclusion that reinsurance quality depends on governance throughout the contract life.

Pricing, Reserves, Cash Flows, and Capital

A technical reinsurance price begins with the probability distribution of contract-eligible recoveries, not simply the distribution of primary claims. Let P represent ceded premium, R_t the recoverable amount due at time t under a defined treaty, and v_t an appropriate discount factor. A simple expected present value of claim recovery is:

EPV⁡(R)=E[∑tvtRt].\operatorname{EPV}(R)=E\left[\sum_t v_tR_t\right].

A quoted premium may then incorporate administration and brokerage costs, acquisition expenses, profit targets, risk margins, capital costs, reinstatement premiums, experience adjustments, and commercial terms. The decomposition is conceptual; no universal pricing formula fixes these elements in all markets. The analyst should make explicit whether the estimate is gross or net of commission, whether cash flows are discounted, and whose capital cost it represents.

For specific excess coverage, expected ceded recovery is the integral of the loss tail above attachment up to limit. If C is a continuous eligible per-person claim amount, the familiar uncapped stop-loss premium component is E[(C−A)+]E[(C-A)_+], equal to ∫A∞P(C>x) dx\int_A^\infty P(C>x)\,dx under standard integrability conditions. With a limit L, the expectation becomes ∫AA+LP(C>x) dx\int_A^{A+L}P(C>x)\,dx. This identity makes the statistical problem transparent: the upper survival curve, which is often estimated from sparse and evolving data, directly governs the technical layer cost. Medical inflation can alter that survival curve relative to a fixed nominal attachment even if clinical incidence is unchanged.

Aggregate excess coverage depends on the distribution of total eligible S, including correlations among claimants. A compound distribution built from independent frequency and severity assumptions can serve as an initial approximation, but correlated shock scenarios must be added where appropriate. Simple deterministic loading of the expected loss is insufficient to characterize the possibility of layer exhaustion. Appropriate stress tests should vary loss frequency, high-cost severity, provider reimbursement levels, reporting delay, health-event dependence, and loss development together, rather than moving only a single average trend parameter.

Disability pricing adds cash-flow duration. An insurer may cede benefits expected over many future years; modeling therefore requires transition rates by claim age and duration, benefit amounts, indexation, lapse/termination provisions, offsets, discount curves, and reinsurance settlement rules. A present-value transfer that appears well priced at claim onset can produce adverse later development if actual recovery rates weaken. Long-term treaties may have experience-account or profit-sharing terms that modify ultimate premium and risk transfer. The analysis must distinguish material future biometric risk from arrangements whose economic substance is principally financing.

Reserve measurement requires matching outstanding gross liability to ceded contract coverage. A recoverable balance reflects future contingent payments from the reinsurer under actual terms, not a simplistic fraction of the total gross reserve unless the treaty truly prescribes that fraction for those obligations. Outstanding case reserves, incurred-but-not-reported claims, disabled-life reserves, premium provisions, and related expenses may interact differently with reinsurance. Variance in claim emergence and settlement can cause discrepancies even when ultimate nominal recovery is unchanged. The IFRS 17 standard overview identifies its insurance-contract measurement and reporting framework; the foundation's reinsurance contracts held guide explains why held reinsurance is measured and presented separately from underlying contracts.

Counterparty nonperformance reduces realizable protection. As a stylized approximation, if expected discounted contractual recovery is $10 million, the probability of default on the relevant settlement horizon is 2%, and loss given default is 60%, then an independence-based expected credit reduction equals $10 million × 0.02 × 0.60, or $120,000. The approximation is deliberately simplified: default probability and recovery size can be correlated, collateral may reduce exposure, contract timing matters, and the appropriate regulatory valuation may follow more detailed rules. The Solvency II Article 81 requirements expressly require timing differences between direct payments and reinsurance recoveries and an adjustment for expected counterparty default losses.

The capital effect is more intricate than ceded premium or balance-sheet recoverables. A treaty may lower modeled health or disability underwriting risk while increasing counterparty concentration and liquidity exposure. Capital credit depends on the applicable supervisory framework, the enforceability and substance of risk transfer, the identity and collateral of the reinsurer, and the calculation's risk classification. Under Canada's LICAT reinsurance credit provisions, recognition has specific conditions for registered and unregistered arrangements. No claim is made that those Canadian criteria automatically control non-Canadian cedents.

Liquidity deserves its own assessment. Consider an insurer owing $12 million in eligible primary payments in one month but contractually receiving $8 million of ceded reimbursement only after two subsequent months. Even if the reinsurer ultimately pays in full, the cedent must bridge the timing gap. A health event may make immediate cash calls large at the same time that health providers accelerate invoices or reinsurer adjudication slows. A balance-sheet reduction in expected ultimate loss is therefore not equivalent to immediate claims-paying capacity. Collateral availability, cash settlement clauses, and contingency liquidity facilities should be assessed using dated scenarios.

The final economic appraisal considers retained expected loss, retained adverse-tail metrics, claims-paying liquidity, administrative effort, premium and commissions, counterparty quality, and contract duration. A layer that is attractive on expected ceded claims may be poor value at the margin if its attachment leaves the portfolio's dominant accumulation untouched or its limit exhausts too readily. Conversely, a costly-looking treaty may provide valuable protection against a low-frequency concentration beyond ordinary capital tolerances. The appraisal reports the trade-offs explicitly rather than asserting that maximum cession or minimum premium is inherently optimal.

International Regulation, Public Programs, and Accounting

The International Association of Insurance Supervisors' ICP 13 on reinsurance supplies the international supervisory reference point. It addresses how supervisors expect insurers to manage their use of reinsurance and other forms of risk transfer and how reinsurers are supervised. Its 2024 revision belongs to the insurance core framework, rather than functioning as a directly enforceable statute in every jurisdiction. The IAIS (2026) ICP 13 peer-review account reports participation by 80 jurisdictions and a high level of observance, with 85% receiving Observed or Largely Observed assessments. This is evidence about supervisory frameworks and their implementation, not direct evidence about the loss ratio of disability or medical reinsurance treaties.

Within the European Union's Solvency II system, health underwriting risk is subdivided into non-life-technique health, life-technique health, and health catastrophe risk. The applicable classification depends on the obligation's technical basis. The health underwriting module specification and health disability-morbidity provision demonstrate that health and disability claims do not occupy one uniform regulatory risk bucket. Capital changes must be evaluated in the actual line and model relevant to the cedent, using the applicable legal version and undertaking-specific exposure.

EU Solvency II also requires a separate treatment of reinsurance recoverables and related expected credit losses. Article 81 on recoverables addresses settlement timing and counterparty default; the delegated recoverables provisions address consistency with contract boundaries and the calculation of amounts recoverable. The combination is important: an insurer cannot substitute a gross premium- or reserve-based assumption for contractual evidence of what the reinsurance actually covers. Detailed recognition conditions and later amendments must be checked for the relevant date of financial reporting or capital assessment.

In the United States, the NAIC reinsurance regulatory overview and its discussion of the credit-for-reinsurance models explain the supervisory framework and reciprocal-jurisdiction arrangements. State adoption, eligibility, and the actual license and contract remain material. The NAIC's (2015) employer stop-loss analysis shows why a self-funded employer's stop-loss policy differs from reinsurance bought by a licensed primary carrier. The Department of Labor stop-loss release further specifies how federal ERISA principles interact with state authority over insurance contracts sold to employers. A scientific comparison must identify the insured buyer, contract, regulatory layer, and relevant benefit obligations.

U.S. public risk-sharing mechanisms need separate categorization. HHS-operated risk adjustment includes a high-cost risk pool mechanism for certain regulated commercial markets, with prescribed data reporting and audit. CMS also maintains documentation for Section 1332 state innovation waivers, some of which support state reinsurance structures. These arrangements can reduce high-cost exposure or support insurance-market stability, but their authority, fiscal payments, participants, and covered losses differ from ordinary cedent–reinsurer treaty relationships. The word “reinsurance” in a program's name does not settle the legal nature of its mechanism.

Canada's OSFI (2025) reinsurance governance guideline treats reinsurance as a strategic risk-management tool requiring attention to legal enforceability, credit concentration, due diligence, cession limits, and liquidity. The LICAT credit provisions explain relevant life-insurer treatment, including the distinction between registered and unregistered reinsurance. These documents show how an international comparison must preserve jurisdictional differences rather than import a uniform collateral or solvency assumption.

IFRS 17 provides the major international financial-reporting perspective for entities within its scope. The IFRS 17 standard applies from annual reporting periods beginning on or after January 1, 2023, subject to specified early-application conditions. The IFRS Foundation's reinsurance held educational guide sets out how insurance contracts issued and reinsurance contracts held are treated as distinct accounting groups, including separate presentation. The guide was developed against the 2017 standard and expressly notes subsequent amendments; it should be read with the current standard rather than mistaken for an exhaustive as-amended rulebook. This matters particularly for loss-recovery components, cash-flow estimation, discounting, and adverse counterparty events.

Professional actuarial guidance can be decisive even where it is not a statute. ASOP 11 addresses actuarial services concerning financial reports reflecting life, annuity, or health-related reinsurance and similar transfer programs, including defined applications to employer stop-loss. ASOP 23 concerns data quality, while ASOP 56 covers actuarial modeling. Those standards govern covered practitioners in their professional setting. The content can inform comparative scientific method elsewhere, but compliance obligations cannot be presumed universal.

Public-health financing mechanisms confirm the importance of purpose and trigger design. The World Bank's Pandemic Emergency Financing Facility pooled support for outbreak response and used an insurance-linked window with parametric activation. Its documented COVID-19 disbursements were emergency public financing rather than claims reimbursement to private health insurers. It provides evidence of a health-linked catastrophe transfer mechanism, while illustrating that trigger-based relief and indemnity under a medical insurance treaty answer different questions.

Integrated Comparative Findings and Research Method

The common scientific problem is to measure how contingent medical or disability obligations propagate through insurance contracts, reinsurance layers, settlement processes, and capital arrangements. Comparison is more reliable when each layer is measured on the same loss definition and temporal horizon. A gross claim distribution estimates covered obligations under the direct policy. Applying treaty terms produces contractual ceded recoveries. Adding settlement timing and counterparty conditions produces realized or realizable cash flows. Regulatory and accounting frameworks then determine whether and how the rights are recognized for particular reporting purposes.

A compact comparative reading follows.

Protection structure Primary reinsured unit Main benefit Characteristic unprotected exposure
Quota share Agreed proportional portfolio obligations Broad sharing of eligible losses and growth Retained percentage, exclusions, credit exposure
Specific medical excess Eligible annual or episode loss of a person Very high-cost individual claimants Loss below attachment, uncovered expense, exhaustion
Aggregate health stop-loss Eligible portfolio loss over a specified period Frequency, severity and correlated annual deterioration Amount below attachment, cap exhaustion, basis mismatch
Disability proportional participation Eligible incurred disability benefits Sharing incidence and benefit-payment risk Claim duration excluded by treaty, timing, default
Disability excess or block structure Defined per-claim liability or reserve exposure Severe duration or concentration of liabilities Definitions of disability, recurrences, valuation basis
Public risk pool or parametric facility Statutorily defined issuer expense or disease trigger Market stabilization or emergency financing Differences between statutory triggers and insurer claims

The table compares economic functions. It makes no assertion that every structure is offered in each jurisdiction or that a governmental payment arrangement is legally a reinsurance treaty.

Three results emerge from the formal illustrations. First, specific reinsurance is driven by the claim distribution above attachment; raising an attachment by a small amount can affect expected recovery substantially if many claims lie near it. Second, aggregate reinsurance responds to joint loss development and therefore depends on correlation, not simply on individual severity. Third, disability reinsurance exposed to future benefit duration can deteriorate even while new-claim incidence remains unchanged. These results follow from contractual loss functions and elementary stochastic mathematics. Their empirical size must be estimated separately for each portfolio.

Further conclusions depend on the institutional evidence. Surveyed U.S. medical expenditure concentration establishes a reason to examine individual medical outliers, but not the recovery percentage of a particular treaty. Group LTD experience studies show both the scale and sampling boundaries of morbidity data; published aggregates cannot replace an insurer's exposure history and claims adjudication rules. IAIS, Solvency II, OSFI, NAIC, and IFRS materials establish distinct layers of supervisory and reporting requirements, each conditional on jurisdiction and applicable basis. No single regulatory framework turns a contractual payment limit into an unconditional capital asset.

The research used documented primary regulations, institutional studies, official statistics, and a bounded sample of peer-reviewed health-economics and actuarial literature. It checked title, identity, dated applicability, and the substantive propositions relied upon in the cited official online materials; the 2015 NAIC stop-loss and 2025 SOA incidence PDFs were examined with page-level content. Some referenced educational guides describe earlier versions of standards, and detailed industry datasets behind the SOA studies require access not used in this investigation. No proprietary claim records, nonpublic treaty documents, individual patient data, or jurisdiction-by-jurisdiction statutory audit were acquired. All numerical illustrations are author-derived and can be checked independently from the equations and stated assumptions.

A reinsurance program should therefore be evaluated by a sequence of questions: What exactly is the direct obligation? Which loss is eligible for cession? What event and accounting period trigger the treaty? Which payer must settle which cash flow at which date? How does dependence affect aggregate results? What happens when the limit exhausts or a counterparty defaults? What part of the apparent reduction is recognized in capital and reporting? This is an ordered research method for the present comparison, not a newly named doctrine or proprietary actuarial model.

The original analytical contribution of this dossier is the integrated comparison across medical severity, disability duration, provider and event accumulation, contractual trigger design, settlement timing, and legal-financial recognition. It identifies where importing a measure from one domain into another produces false inferences: medical expenditure shares into ceded loss ratios, incidence into disability duration, nominal limits into collectible liquidity, professional standards into foreign statutes, and public high-cost risk pools into private treaty accounting. These are source-grounded analytical conclusions and domain distinctions. Claims of priority or a new universal actuarial theory would exceed the evidence and are not made.

Conclusion

Health and disability reinsurance programs redistribute the financial cost of insured medical expenses, incapacity, and correlated morbidity through contractually specified risk transfer. Their scientific analysis requires separate treatment of high-cost individual claims, aggregate portfolio deterioration, disability incidence, benefit duration, and the conditions governing actual recoveries. The unit of covered risk, attachment threshold, treaty period, and definition of eligible loss determine the mathematical recovery function; medical and disability experience determine the distribution of its inputs; contractual performance, liquidity, and regulation determine its financial significance.

Specific excess-of-loss protection is especially responsive to the upper tail of claimant severity, aggregate stop-loss to frequency and dependence across the portfolio, proportional arrangements to broad sharing, and duration-sensitive structures to the future cost of open disability claims. A reinsurance program that appears sufficient against one risk can remain exposed to another: long duration despite stable incidence, broad medical inflation despite few extreme claimants, correlated losses despite a large insured population, or delayed reimbursement despite a substantial ultimate recovery. Actual treaty wording and coherent loss-data definitions determine which protection is realized.

The international evidence confirms a layered institutional environment. The IAIS supplies cross-jurisdiction supervisory principles; Solvency II assigns health underwriting and counterparty risks to defined capital treatment; IFRS 17 requires distinct accounting for insurance contracts issued and reinsurance held; North American regulatory documents distinguish private reinsurance, employer stop-loss, and government risk-sharing structures. These arrangements solve related financing problems through different legal obligations. Comparison becomes scientifically meaningful when those differences are preserved.

The resulting conclusion is that reinsurance effectiveness is a property of the complete relationship among underlying benefits, statistical exposures, contract mechanics, counterparty performance, and legal recognition. Measuring only ceded premiums or nominal treaty limits supplies an incomplete view. Transparent claim-level and portfolio-level calculations, disability transition analysis, accumulation scenarios, reliable source documentation, and dated cash-flow stress tests produce a stronger basis for understanding program design. This monograph establishes that comparative analysis without creating a new theory, proprietary system, or unsupported empirical estimate.

References

  • Actuarial Standards Board. (2016). Actuarial Standard of Practice No. 23: Data Quality (revised December 2016). Official standard.

  • Actuarial Standards Board. (2019). Actuarial Standard of Practice No. 56: Modeling (revised December 2019; effective October 1, 2020). Official standard.

  • Actuarial Standards Board. (2021). Actuarial Standard of Practice No. 11: Treatment of Reinsurance or Similar Risk Transfer Programs Involving Life Insurance, Annuities, or Health Benefit Plans in Financial Reports (effective December 1, 2022). Official standard.

  • Arrow, K. J. (1963). Uncertainty and the welfare economics of medical care. American Economic Review, 53(5), 941–973. Publisher-hosted article.

  • Centers for Medicare & Medicaid Services. (2026, page updated July 28). CCIIO examinations, audits and reviews of issuers: Issuer resources (including high-cost risk-pool audits). CMS official resource.

  • Centers for Medicare & Medicaid Services. (2026, accessed November 14). Section 1332: State innovation waivers. CMS official documentation.

  • Cutler, D. M., & Zeckhauser, R. J. (2000). The anatomy of health insurance. In A. J. Culyer & J. P. Newhouse (Eds.), Handbook of Health Economics (Vol. 1A, pp. 563–643). Elsevier. DOI: 10.1016/S1574-0064(00)80170-5. Publisher record.

  • European Insurance and Occupational Pensions Authority. (current online rulebook, accessed November 14, 2026). Solvency II Directive 2009/138/EC, Article 81: Recoverables from reinsurance contracts and special purpose vehicles. EIOPA official rulebook.

  • European Insurance and Occupational Pensions Authority. (current online rulebook, accessed November 14, 2026). Delegated Regulation (EU) 2015/35, Article 41: General provisions concerning reinsurance recoverables. EIOPA official rulebook.

  • European Insurance and Occupational Pensions Authority. (current online rulebook, accessed November 14, 2026). Delegated Regulation (EU) 2015/35: Health underwriting risk module. EIOPA official rulebook.

  • European Insurance and Occupational Pensions Authority. (current online rulebook, accessed November 14, 2026). Delegated Regulation (EU) 2015/35: Health disability-morbidity risk sub-module. EIOPA official rulebook.

  • Finkelstein, A., & McGarry, K. (2006). Multiple dimensions of private information: Evidence from the long-term care insurance market. American Economic Review, 96(4), 938–958. DOI: 10.1257/aer.96.4.938. Open full-text article.

  • Haberman, S., & Pitacco, E. (1998). Actuarial Models for Disability Insurance. Chapman & Hall/CRC. ISBN 9780849303890. Publisher book record.

  • Hernandez-Viver, A., & Mitchell, E. M. (2025, March). Concentration of healthcare expenditures and selected characteristics of people with high expenses, United States civilian noninstitutionalized population, 2018–2022. MEPS Statistical Brief No. 560. Agency for Healthcare Research and Quality. AHRQ official study.

  • IFRS Foundation. (2018). IFRS 17: Pocket guide on reinsurance contracts held (educational material reflecting the standard as issued in May 2017; see later amendments). IFRS educational guide.

  • IFRS Foundation. (2026, accessed November 14). IFRS 17 Insurance Contracts (standard overview and current materials). IFRS 17 standard page.

  • International Association of Insurance Supervisors. (2024, December revision). Insurance Core Principles and ComFrame, ICP 13: Reinsurance and other forms of risk transfer. IAIS official ICP 13.

  • International Association of Insurance Supervisors. (2026, April 21). IAIS publishes aggregate report on the peer review process on ICP 13 (Reinsurance and other forms of risk transfer). IAIS peer-review release.

  • National Association of Insurance Commissioners. (2015). Stop Loss Insurance, Self-Funding and the ACA. NAIC white paper. Official NAIC white paper.

  • National Association of Insurance Commissioners. (2025, October 24). Reinsurance. Insurance Topics. Official NAIC overview.

  • Office of the Superintendent of Financial Institutions Canada. (2025, January 1). Sound Reinsurance Practices and Procedures: Guideline. OSFI official guideline.

  • Office of the Superintendent of Financial Institutions Canada. (2025, January 1). Life Insurance Capital Adequacy Test (2025), Chapter 10: Credit for Reinsurance (with associated regulatory notice in 2025). OSFI LICAT Chapter 10.

  • Society of Actuaries. (2013). Group Long-Term Disability 2012 Valuation Table and associated work group materials. SOA valuation resources.

  • Society of Actuaries Research Institute & LIMRA. (2025, September). 2015–2022 Group Long-Term Disability Incidence Study Report. SOA incidence study.

  • Society of Actuaries Research Institute & LIMRA. (2026, September 30). 2015–2024 Group Long-Term Disability Termination Experience Study. SOA termination study.

  • U.S. Department of Labor, Employee Benefits Security Administration. (2014, November 6). Technical Release No. 2014-01: Guidance on State Regulation of Stop-Loss Insurance. Official DOL release.

  • World Bank. (2020, April 27; updated to record facility closure). Fact Sheet: Pandemic Emergency Financing Facility. World Bank official fact sheet.