Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Long-Term Care Reserve Adequacy and Assumption Updating

Abstract. Long-term care (LTC) insurance liabilities are unusually sensitive to changes in assumptions because the elapsed time between policy issue and claim settlement can extend over decades. The reserve valuation must integrate the probability of entering covered care, the distribution of care intensity and claim duration, mortality, recovery, lapse, contractual premium rights, benefit inflation, expenses, discounting, and the performance of supporting assets. This monograph examines how insurers can determine whether reserves are adequate and how emerging experience should alter the assumptions underlying that determination. It reviews actuarial professional standards, U.S. regulatory guidance for LTC asset adequacy, published intercompany experience, IFRS 17, Solvency II, and international evidence about the cost of long-term care. Standard cash-flow and state-transition representations are used to establish a common analytical vocabulary, while a transparent hypothetical sensitivity example distinguishes individual and interacting assumption effects. The central finding is that reserve adequacy is a framework-specific, valuation-date judgment derived from contract obligations, credible and comparable experience, documented uncertainty, and, where required, tests of supporting asset cash flows. Aggregate industry observations supply valuable benchmarks but cannot replace insurer-specific cohort analysis and validation. Original conclusions concern the connection between measurement, assumption updating, reporting, and solvency; no new actuarial theory or original empirical estimate is claimed.

Keywords: Long-term care insurance; reserve adequacy; assumption updating; morbidity; claims incidence; claim termination; mortality; voluntary lapse; insurance persistency; actuarial credibility; cash-flow testing; AG 51; asset adequacy; Solvency II; IFRS 17; risk adjustment; policyholder protection; technical provisions.

Introduction

Reserve adequacy in long-term care insurance is a question about future contractual payments and the resources designated to meet them. The characteristic timing of LTC liabilities creates difficulty for any valuation. Coverage may begin at middle age while the principal claims experience emerges decades later, when the insured population, services, care institutions, treatment practices, wages, investment environment, and market regulation have evolved. A projection calibrated at policy issue is exposed to systematic divergence as the block matures. The purpose of an assumption update is to measure that divergence and determine its consequences for the current estimate of obligations.

Long-term care need, policy eligibility, claim incidence, benefit payments, and reserve liabilities are related but distinct observational objects. Care may be provided informally without a claim. An eligible claim may be delayed by an elimination period. A payment may be constrained by a fixed daily limit, a pool of lifetime benefits, a qualifying-care setting, or other operative policy terms. The OECD's comparative research on care affordability establishes the breadth of national variation in service costs and publicly funded benefits; it does not estimate the insured claim rate of any private carrier. The liability model must translate the precise insured promise rather than substitute population-level service demand for contractually covered payments.

The research asks three connected questions. What does a sufficient LTC reserve measure under a defined legal and accounting basis? Which data and statistical methods support changes to morbidity, utilization, termination, lapse, mortality, and financial assumptions? How do these changes affect statutory reserve and asset adequacy analysis, IFRS 17 results, and prudential technical provisions? The answers depend on interpreting original evidence and distinguishing current scientific findings from prospective research programs.

The methodological foundation combines documentary analysis of primary regulations and professional standards, critical review of Society of Actuaries (SOA) and LIMRA intercompany experience, established actuarial projection methods, and an algebraically reproducible sensitivity illustration. The research employs U.S. evidence as an important empirical source while keeping the scientific scope international. No confidential insurer data, actual company financial statements, independently estimated hazard model, or original causal experiment enter the study. Each quantitative illustration is identified as hypothetical.

The following chapters progress from the definition of the insured obligation through its measurement, assumptions, financial reporting, and scientific validation. The results address the substance of an actuarial reserve judgment and the conditions required for its reproducibility.

Conceptual Foundations and Historical Development

The meaning of an adequate reserve

A reserve is a liability measure established under an applicable accounting or regulatory framework. An adequacy assessment asks whether that measure and, in contexts requiring asset testing, the cash flows from assets supporting it are sufficient under the relevant rules and test conditions. Different frameworks use different discount rates, boundaries, margins, and recognition of profit. Therefore, an identical portfolio can produce several legitimate valuations, each serving its defined reporting purpose. A serious analysis begins by stating the purpose and legal or accounting basis before discussing whether the result is adequate.

Long-term care contracts may promise fixed cash indemnities, reimbursement of eligible expenditures, nursing-facility benefits, qualifying home or community services, inflation protection, or combinations of benefits. A material reserve forecast depends on the exact definitions of covered dependency and the monetary limits of these promises. The adopted ASOP No. 18 on Long-Term Care discusses the assumptions used for long-term care analysis, sensitivity testing, and monitoring of emerging experience. The combination of product design and long duration demands far more than a single average claim cost.

A valuation must identify its population: policies in force at the valuation date, currently open claims, eligible renewal premiums, remaining benefit pools, reinsurance and other contractual offsets, and incurred but unpaid obligations. Under a cash-flow test it must also identify the supporting assets and their expected investment cash flows. These elements must be reconciled to avoid counting a future claim in both an incurred-claim liability and a prospective coverage liability.

Different clocks govern different risks

An LTC policy has multiple durations: time since issuance, age attained by the policyholder, calendar time, time until first qualifying claim, duration since claim onset, time spent in a care setting, and remaining benefit duration. These clocks are not equivalent. The claim onset hazard for a newly underwritten policyholder aged 70 may differ from the hazard for a persistently insured person of the same age whose policy was issued decades earlier. A claimant receiving covered home care for one month faces different competing termination risks than a claimant receiving institutional care for three years.

The SOA's 2000–2016 LTC experience report obtained material from 18 companies, described as representing 80% of U.S. LTC earned premiums in 2016. Its incidence aggregate included 61,323,281 total exposure units and 620,591 claims, or approximately 1.012% on that specified aggregate basis. The report distinguishes active-lives and total-lives denominators and documents rules for episode aggregation. Those aggregate numbers characterize the study, not a universally applicable incidence assumption. They cannot be projected for decades without standardizing for age, policy duration, contract design, and the insured population.

Incidence, termination, and utilization

Three actuarial processes describe different parts of LTC morbidity. Claim incidence measures the entry into covered claim status. Claim termination measures exit from claim status through death, recovery, benefit exhaustion, contractual termination, and related causes. Utilization measures the intensity and type of benefits consumed while claim status persists. A deterioration in any of these components can increase expected future payments. For example, unchanged onset rates with slower recovery may increase the mean duration and present value of claims; unchanged incidence and termination may still produce greater payments if higher-cost covered services become more prevalent.

The SOA's LTC basic experience tables separately modeled incidence, claim termination, and claim utilization using historical data from 22 insurers for 2000–2011. These studies support actuarial decomposition, but they also describe experience from a particular period and product population. The relevance of their parameters to newer policies depends on differences in underwriting, benefit triggers, inflation protection, claim management, and insured selection.

Lapse, death, and insurance selection

In-force policies terminate for different reasons. Voluntary lapses, mortality, and transitions out of premium-paying status affect the number of lives exposed to later claims and the future premiums available to finance those claims. The SOA/LIMRA policy persistency study analyzes lapse, mortality, and total termination activity from 18 participating companies during 2000–2016. A separate mortality and lapse valuation study proposes valuation bases and documents the research used to derive them. Such studies provide empirical and technical support; adoption as a regulatory minimum is determined under the applicable rule and jurisdiction.

A decline in voluntary lapse frequently increases future LTC claims because more policies survive into high-claim ages. The net present value effect also depends on the premiums retained, the benefit structure, and the probability of future claims. The correct calculation is a multi-period cash-flow result rather than an assumption that lower lapse always produces an identical increase in all reserve measures. Mortality improvement likewise affects the timing and composition of claims differently for healthy insureds and current claimants.

Selection can complicate experience comparisons. Finkelstein and McGarry's study of private information in LTC finds dimensions of information about risk and insurance preferences that can offset in observed market correlations. The absence of a simple positive association between coverage and subsequent claims does not demonstrate that insurance selection is absent. A company's issue-era underwriting rules, distribution channels, and subsequent policyholder behavior can all change the exposure mix and the validity of historical industry assumptions.

The development of regulatory adequacy testing

Statutory minimum valuation and long-horizon asset adequacy testing address related issues but employ different tests. The NAIC's Actuarial Guideline LI, known as AG 51, sets out specific concerns regarding testing of LTC reserves and the support for assumptions concerning morbidity, mortality, lapse, investment returns, and rate increases. Its discussion refers to the Health Insurance Reserves Model Regulation, VM-25, and asset adequacy under VM-30. An analysis must determine which version applies to the entity and valuation date. The guideline provides a documentary foundation for the distinction between a minimum formula and an assessment of assets supporting a material LTC block.

The ASOP No. 22 asset-adequacy standard calls for an appropriate test horizon, reasoned selection of methods, scenario analysis, review of management actions, and supporting documentation. An appointed actuary's opinion is a professional judgment under applicable law, supported by documented tests. It is not reducible to the ratio of a published claim count and total policy exposure.

Reserve Basis and Liability Duration

A prospective financial representation

Define the economic net present value at valuation time t for a specified group of existing contracts as

Vt=E⁡t[∑u>tDt,u(Bu+Eu−Pu−Ru)].V_t=\operatorname{E}_t\left[\sum_{u>t}D_{t,u}\left(B_u+E_u-P_u-R_u\right)\right].

Here B denotes future covered benefit payments, E qualifying expenses, P qualifying future premium receipts, R permissible recoveries or offsets, and D discount factors appropriate to the defined purpose. The expectation is conditional on information available at t. This expression is an explanatory net-cash-flow building block, not a substitute for any prescribed statutory formula. A regulatory or accounting liability may add specific margins, adjust contract boundaries, reflect reinsurance separately, prescribe assumptions, or recognize future profit through another component.

The expected value depends on the probabilities of transitions among insurance states and the cash flows conditional on those states. An admissible state set may contain active without claim, home-care claim, residential-care claim, recovered or inactive coverage, lapsed, benefit exhausted, and deceased. More detailed states may be necessary to represent cognitive impairment, contractual triggers, waiver of premium, claim severity, and remaining available benefits. The number of states should be determined by predictive and payment relevance and the credibility of the data.

For a discrete-time representation, write

πt+1=πtPt,∑jPij,t=1,\boldsymbol{\pi}_{t+1}=\boldsymbol{\pi}_t\mathbf P_t,\qquad \sum_jP_{ij,t}=1,

where the row vector contains probabilities of occupying each state and the matrix contains conditional transition probabilities. The payment expectation in a period follows from state-specific contractual amounts and exposures. A Markov model is useful when the current modeled state and explanatory variables adequately capture relevant transition history. If claim duration materially affects the probability of recovery or death, an expanded-duration or semi-Markov structure is more appropriate. The equations summarize established actuarial modeling principles, not an original theoretical invention.

Claim duration, competing exits, and censoring

Claims terminate through competing causes. A claim-duration model should distinguish recovery, death, exhaustion of insured benefits, and other contractual exits. The distribution of benefit payments is influenced by both the time of exit and the care setting before exit. Censoring occurs when the study ends while a claim is still open; ignoring such claims biases duration inference. A dataset of closed claims cannot automatically estimate the expected remaining life of open claims.

A duration hazard should be conditioned on survival in claim status to the start of the period. Competing exit hazards affect the survival probability together, and expected benefits can be summed over future periods weighted by the probability of continued covered status. For an existing claimant, the appropriate projection also reflects benefit units already used, remaining maximums, care setting, and any qualifying conditions. For a currently healthy insured, it integrates both survival in eligible policy status and future claim onset.

Utilization, inflation protection, and benefit ceilings

The contractual benefit calculation transforms the need for care into the insurer's payment. An indemnity policy with a fixed daily benefit may pay the same amount regardless of the external provider's full cost. A reimbursement policy may pay a variable amount up to a specified ceiling. Policies with compound inflation protection, pooled lifetime limits, or different institutional and home-care rates transmit care-cost developments differently to the insurer. A projected cost increase may be fully borne by the insured person after a cap is reached, or may enlarge the insurer's expected payments when coverage remains open-ended within contractual terms.

This distinction is important in international analysis. OECD evidence on care cost-sharing documents large variation in publicly financed care and private affordability. Such evidence informs likely service environments and scenarios, but it does not establish an LTC insurer's benefit liability until the covered-care contract and public/private coordination are specified.

Separate incurred and future-coverage obligations

A robust reserve architecture distinguishes payments arising from incurred claims and those that may arise from coverage still to be provided. At a valuation date, an open claim has a conditional remaining payment distribution. A policy without an open claim can still have a large prospective liability at later attained ages. Costs associated with care already incurred but not fully reported should be allocated consistently with the reporting basis and must not be duplicated across categories.

The relevant premium stream can be restricted by contractual boundaries, lapse, mortality, waiver of premium, or changes in policyholder status. A policy that ceases premium collection during claim payment may have substantial claim outflows without associated renewal premium receipts. Reinsurance recoverables may have their own default and timing assumptions. A technically plausible projection can still misstate reserves if benefit and premium timing are mismatched.

Discount rates, investment income, and liquidity

Discounting affects long-duration liabilities particularly strongly. The present value of one fixed payment C at duration n under a constant annual effective rate r equals C(1+r)^{-n}. Its sensitivity to a small rate change grows roughly with n. Actual LTC cash flows consist of a changing distribution of premiums, benefits, expenses, and claim states; the simple derivative illustrates duration exposure without estimating the full portfolio's economic sensitivity.

Asset adequacy tests introduce actual investment behavior. Reinvestment rates, asset duration, credit losses, liquidity, embedded options, and the availability of cash when claims fall due affect the ability of supporting assets to fund the liability. The revised ASOP No. 7 cash-flow standard, effective in June 2026, provides professional guidance for life or health insurer cash-flow analysis. Asset liquidity and future yield cannot be established merely by choosing a discount rate for the liability.

The reserve movement bridge

Changes in reported reserve from one period to another should be reconciled among expected run-off, claim experience, revised exposure, incidence, termination, utilization, lapse, mortality, expenses, discount curves, model changes, and changes in management-action assumptions. A useful reconciliation starts from the prior valuation's frozen basis and applies successive changes while documenting ordering and any interaction residual. When assumptions interact, attributing a joint change solely to one factor overstates its explanatory power.

The actuarial committee should be able to trace a proposed reserve revision to a specific evidence change, the parameter response, and the modeled cash-flow effect. The test of credibility is not merely that the revised reserve is larger or smaller; it is that the method explains why the changed estimate follows from the observed facts and applicable requirements.

Assumption Revision and Uncertainty

Identifying the source of a deviation

Observed paid claims are a product of exposure, eligibility, incidence, duration, utilization, and payment rules. An increase may result from more insureds attaining older ages, earlier claims reporting, a change in claim definition, longer covered episodes, higher care intensity, contractual benefit growth, or a genuine increase in conditional care incidence. An update based only on aggregate paid dollars will mix these causes. The first scientific task is to isolate comparable measures and test their stability after adjusting for population mix.

The ASOP No. 18 experience-monitoring guidance supports a process in which emerging experience is compared with prior assumptions and material differences investigated. Each model parameter should have an identifiable estimand, exposure denominator, observation window, data owner, and validation diagnostic.

Incidence estimation

A crude incidence estimator for a defined cell is \hat q=N/E, with N qualifying new claim episodes and E the corresponding exposure. Its interpretation depends on whether E includes all lives in force, only nonclaiming lives, policy years, or an adjusted at-risk period. The SOA's aggregate incidence methodology expressly distinguishes active-lives exposure from total-lives exposure, and establishes a method for counting claims close together in time. Two studies using different definitions can yield different incidence estimates for the same underlying policy population.

Rates should be analyzed by relevant attained age, sex where analytically and legally appropriate, policy duration, issue cohort, product type, underwriting, and calendar period, subject to credible cell sizes. A Poisson approximation can provide an initial diagnostic for sparse independent events, but unobserved heterogeneity and dependence often require richer variance estimation. Age standardization and duration stratification help separate portfolio maturation from a changing conditional risk rate.

Termination and utilization estimation

A claim-termination assumption must reflect duration on claim and the competing mechanisms of recovery, death, and benefit exhaustion. An analysis that treats every administrative closure as full recovery will misstate expected future claims. A study that observes only settled claims will omit long open episodes and bias estimates of duration. Relevant methods include survival analysis with censoring, cause-specific competing-risk hazards, and multi-state transition models.

Utilization assumptions require the measured quantity: days paid, hours provided, billable services, proportion of a policy maximum, or amount per active claim month. A trend in nominal costs should be separated from a trend in units of care and a shift toward more expensive covered settings. Medical inflation, care-wage inflation, and insurer payment escalation are distinct variables when contractual caps intervene.

Mortality and persistence updates

Healthy-life mortality, claimant mortality, voluntary lapse, and total policy termination should be examined separately. The SOA/LIMRA termination studies were designed to distinguish components. The later individual LTC persistency evidence extends the available historical observation period through 2016. Their relationship to a specific current block depends on the company's own lapse behavior and claim mix.

Lower lapse commonly increases later claim exposure, while also retaining premiums before the claim. A credible update therefore calculates both sides. If a carrier raises premiums, different insureds may lapse or reduce benefits at different rates; those who remain can be selected on health, risk perception, or value placed on coverage. Future persistence in a re-rated block should not be assumed to follow the same pattern as persistence before the action without an empirical argument.

Credibility and reference experience

Company experience may be thin in older ages, long claim durations, or recently introduced benefit designs. Statistical credibility is needed to combine such observations with external benchmarks. A schematic estimate is \hat\theta=w\hat\theta_{\rm company}+(1-w)\hat\theta_{\rm reference}, where the weight must be justified by the statistical model, information content, and target population. This expression illustrates a family of established credibility methods; no universal fixed weight is asserted.

The external reference must measure the same risk. A lapse benchmark cannot directly substitute for a mortality hazard, and an industry claim-termination table from 2000–2011 cannot be used without review for later contracts with different claim triggers. The SOA LTC tables and development report exemplify how actuarial experience can be stratified and modeled. The insurer must document the mapping from historical industry factors to its own contractual population, as well as the uncertainty introduced by that mapping.

Data vintage and the status of new studies

The SOA/LIMRA 2000–2023 LTC data solicitation called for stand-alone LTC data, with hybrid and combination products excluded from that research scope. The institutions' August 2025 study launch describes a research initiative, not already completed public incidence and termination tables. A later study cannot be treated as a quantitative authority before its results and methods have actually been published and verified.

An assumptions register must record the dates covered by underlying observations, the publication date of each external study, the internal extraction and transformation dates, and the model approval date. A document published recently can still analyze old policy years; a historical observation can remain informative while requiring explicit adjustment. Transparency about the time covered is part of statistical credibility.

Expected values, margins, and risk adjustments

A revised central expectation reflects the current best supported estimate under its framework. Additional adverse-deviation margins, accounting risk adjustments, prudential risk margins, or scenario-based reserves have distinct meanings. Their purposes should be separated to avoid double counting the same adverse exposure. Holding an unchanged margin does not justify withholding a material update to the expected value, and updating the expected value does not automatically determine the proper margin.

The ASOP No. 22 guidance on assumption changes calls attention to changes in assumptions, models, and methods in asset adequacy testing. An impact assessment should show what changed, why it changed, and how materially it affects the cash-flow result. Statistical significance and financial materiality need separate evaluation: a small but persistent claim-duration discrepancy can be financially large in a mature block.

Joint sensitivity analysis: reproducible illustration

Consider an explicitly hypothetical one-period example with 1,000 active policies, a 2% chance of one covered incident in the interval, a conditional discounted claim cost of $80,000, future discounted premiums of $500 per policy, and aggregate expenses of $100,000. The illustrated expected net outflow is 1,000 × 0.02 × $80,000 − $500,000 + $100,000 = $1,200,000. The conditional claim cost already incorporates discounting; the exercise intentionally omits subsequent-year transitions and is not a statutory reserve calculation.

Now raise the incidence probability to 2.5% and conditional cost to $90,000, holding other terms fixed. The result is 1,000 × 0.025 × $90,000 − $500,000 + $100,000 = $1,850,000. The increase is $650,000, or approximately 54.2%. Incidence alone adds $400,000, cost severity alone at original incidence adds $200,000, and the interaction adds $50,000. The arithmetic is internally reconcilable and illustrates the importance of joint stress. It supplies no estimate of a real insurer's reserve or probability distribution.

More realistic scenarios examine combinations of lower lapse, longer claims, different covered settings, longevity change, rising care wages, lower asset yields, and delayed rate approvals. Each scenario should distinguish its evidentiary status: an empirically calibrated distribution, a plausible deterministic adverse path, or a sensitivity deliberately constructed for governance. None should be mislabeled as a calibrated percentile without statistical support.

Assumption governance

A defensible governance cycle contains a frozen prior basis, a reconciled current dataset, diagnostics by homogeneous risk cell, documented proposed changes, financial attribution, independent review, approval by the responsible process, and an archive of the resulting assumptions and model versions. Material subsequent events may require review outside a routine annual cycle. The ASOP No. 23 data-quality standard supports scrutiny of the data used; the ASOP No. 56 modeling standard supports evaluation of the method and its intended purpose.

Financial Reporting and Solvency

U.S. statutory and asset-adequacy context

In the United States, the relevant reserve measure may include minimum liabilities under the applicable health insurance valuation framework and analysis of whether supporting assets can meet future obligations. The NAIC's AG 51 documentation concerns support for assumptions and cash-flow asset adequacy on substantial LTC blocks. Different legal instruments and effective dates must be reconciled. The NAIC model-law catalogue identifies model reserve regulations, but a particular model text is applied to an insurer according to relevant jurisdictional adoption.

The asset adequacy question is not answered merely by observing a prescribed reserve figure. Investment yields, future premiums, claim payments, expenses, and management actions determine the projected sufficiency of resources over the relevant horizon. The ASOP No. 22 professional standard addresses analysis, documentation, scenario interpretation, and management-action assumptions. An unfavorable scenario result requires reasoned evaluation in its full methodological and legal context.

Premium increases are contingent economic actions

Legacy LTC blocks may require premium revisions when initial pricing assumptions prove inadequate. Expected receipts cannot simply assume that every future requested increase will be granted at its proposed level and date. Regulatory approval, contract rights, policyholder elections, benefit reductions, nonforfeiture choices, and changes in lapse behavior can alter cash flows. The NAIC's multistate LTC rate-review framework provides a coordination procedure, not an automatic guarantee of a proposed rate increase.

In asset adequacy testing, the supporting rationale should distinguish approved rate changes, applications in progress, prospective management plans, and purely illustrative alternatives. A modeled rate action must be compatible with the rules applicable to that test and with evidence of implementation feasibility. Its effect on policyholder persistence matters as much as its nominal premium percentage. The actuarial review should quantify both the gross premium change and the behavioral and contractual offsets.

IFRS 17 measurement and current assumptions

The IFRS 17 insurance-contract framework measures groups of insurance contracts using fulfilment cash flows, a risk adjustment for non-financial risk, and a contractual service margin where applicable. Future cash-flow estimates are revised with current information at reporting dates. Changes concerning future insurance services can adjust the contractual service margin subject to the standard's requirements; losses on onerous groups have distinct recognition. An updated LTC incidence estimate consequently need not cause the same immediate profit effect across groups with different CSM balances.

Contract grouping, reporting boundaries, discount rates, incurred-claim classification, and the recognition of future insurance services shape the result. An IFRS liability is not a direct restatement of a U.S. statutory minimum reserve. A rise in expected claims, a discount-curve change, and a revision of a non-financial risk adjustment are distinct drivers even when they occur at the same valuation date. Their accounting consequences should be reported in an explicit movement bridge.

Solvency II best estimate and risk margin

Under Solvency II Article 77, technical provisions generally comprise a probability-weighted present value of future cash flows, using prescribed risk-free term structures, plus a risk margin. The best estimate is based on up-to-date, credible information and realistic assumptions and projects obligations over their lifetime. A prudential best estimate, an IFRS fulfilment-cash-flow estimate, and a U.S. statutory reserve can therefore differ even when based on the same policy portfolio.

EU law also specifies data integrity. Article 82 on data quality requires processes assuring appropriate, complete, and accurate technical-provision data, with approximations in specified circumstances. The delegated rule on approximations clarifies conditions for using methods necessitated by insufficient data. Sparse late-duration observations need sound statistical treatment; poor internal collection cannot automatically justify permanent reliance on rough approximations.

A common comparison basis

A cross-framework reconciliation should begin with the same contract list, evaluation date, currency, and underlying future contingent payments. It should then isolate differences in recognized future premiums, discount curves, incurred-versus-future coverage boundaries, risk adjustment, risk margin, profit recognition, statutory minima, and asset adequacy additions. Only after these bridges can the reported numerical amounts be interpreted.

Element U.S. statutory / asset adequacy IFRS 17 Solvency II
Main purpose Statutory liability and applicable adequacy opinion Financial statement insurance measurement Prudential technical provisions
Core methods Adopted valuation rules and prescribed cash-flow tests Current fulfilment cash flows, risk adjustment, CSM when relevant Best estimate, prescribed discounting, risk margin
Response to evidence Subject to applicable reserve and adequacy rules Reporting-date estimates and service-period treatment Updated realistic expected assumptions
Supporting assets Integral to asset adequacy Separately measured under applicable asset standards Evaluated alongside capital and assets
Management actions Feasibility and approval evidence Standard-specific contract boundaries and cash flows Prudentially justified future actions

The table describes method architecture, not universal equivalence of company figures. An entity-specific analysis must verify the applicable contracts, accounting election, national legislation, regulation and reporting date.

Interaction with public long-term care provision

Private insurance operates within a broader care economy. The OECD's 2024 international care study reports pronounced variation in public support, care prices, and the burden remaining on households. These differences can affect private insured utilization indirectly through provider availability, care-setting choice, and demand. The value of a private benefit still follows the contract's payment terms and coordination rules; population care prevalence must not be treated as insured claim incidence.

Integrated Findings and Research Agenda

The meaning of sufficiency is purpose-specific

Reserve adequacy is a conditional conclusion derived under a particular accounting, statutory or prudential framework at a particular date. The U.S., IFRS 17 and Solvency II rules share an economic concern with future insurance payments while differing in risk treatment, discounting, profit recognition and asset tests. The comparison supports analytical bridges rather than a direct numerical ranking of reserve totals.

Long-horizon claims must be modeled as transitions

Incidence by itself is an insufficient description of LTC exposure. Exit causes, duration dependence, care setting, utilization, lapse and mortality govern the amount and timing of benefits. The SOA studies' separate incidence, termination and utilization components demonstrate why a state-based cash-flow approach is a coherent foundation. Its appropriate sophistication is constrained by the detail and credibility of available observations.

Experience requires cohort and contract comparability

The published U.S. intercompany studies offer valuable denominators, rates, descriptions of data cleansing, and model structures. Their results remain specific to the sampled eras, policy types and definitions. Current blocks require reconciliation of underwriting, issue cohorts, claim triggers, benefit limits and persistence. A study announcement is a proposal for future information, not a substitute for completed data.

Financial assumptions and morbidity interact

Lapse, mortality, rate actions, asset yields and claim severity cannot always be stressed independently. The hypothetical $1.20 million to $1.85 million example identifies a nonzero interaction between incidence and severity. The full LTC portfolio requires richer dependence and time structures; financial reporting should attribute changes consistently rather than bury interaction effects.

A research agenda grounded in testable observations

Empirical work should examine whether cohort-adjusted incidence is changing, how recovery and death hazards vary with claim duration, whether transitions among care settings explain severity inflation, and how approved rate changes affect policyholder behavior and future claim mix. The SOA/LIMRA 2000–2023 proposed experience update illustrates the continuing need for more recent experience, including the limitation that this particular study scope excludes hybrid contracts.

A further priority is an entity-controlled common-portfolio comparison under U.S. statutory reserving, IFRS 17 and Solvency II, with identical basic contingencies but explicit framework-specific reporting bridges. Such research would need confidential longitudinal insured data, contract history, reliable asset information and legally verified assumptions. No claim of empirical discovery is made before such work is performed.

Conclusion

Adequate long-term care reserves depend on the precise translation of enforceable policy promises into future expected cash flows, the quality of data used to estimate transition risks, and the transparent treatment of uncertainty and supporting assets. Long horizons magnify the effects of incidence, duration, persistence, morbidity, care utilization, benefit design and financial conditions. Assumption updating therefore requires defined observations, comparable exposure, statistical judgment, model validation and documented financial attribution.

The evidence establishes that an industry incidence rate is not a portfolio-specific forecast; an administrative claim closure is not necessarily a recovery; a care-cost increase is not automatically the same as an insurer benefit increase; a future requested premium increase is not an approved cash flow; and reserve measures from different reporting frameworks need explicit reconciliation. These distinctions determine the reliability of a long-duration liability calculation and the quality of its solvency interpretation.

The monograph's original contribution is a source-grounded integration of LTC contractual measurement, experience revision, cash-flow projection, regulatory practice, and international financial reporting. Its mathematical forms are established actuarial methods, and its numeric stress is openly hypothetical. The conclusion is substantive and reproducible within the stated evidence boundaries: reserve adequacy rests on a validated connection between measured insured risks, contractual obligations, valuation rules, and the assets and risk provisions supporting them.

References

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

  • Actuarial Standards Board. (2019). Actuarial Standard of Practice No. 56: Modeling. Official text.

  • Actuarial Standards Board. (2021). Actuarial Standard of Practice No. 22: Statements of Actuarial Opinion Based on Asset Adequacy Analysis for Life Insurance, Annuity, or Health Insurance Reserves and Other Liabilities. Revised September 2021; effective June 1, 2022. Official text.

  • Actuarial Standards Board. (2022). Actuarial Standard of Practice No. 18: Long-Term Care. Revised March 2022; effective September 1, 2022. Official text.

  • Actuarial Standards Board. (2025). Actuarial Standard of Practice No. 7: Life or Health Cash Flow Analysis. Revised December 2025; effective June 1, 2026. Official text.

  • European Union. (2009, as amended). Directive 2009/138/EC, Article 77: Calculation of Technical Provisions. EIOPA Single Rulebook. Article 77.

  • European Union. (2009, as amended). Directive 2009/138/EC, Article 82: Data Quality and Application of Approximations. EIOPA Single Rulebook. Article 82.

  • European Union. (2015, as amended). Commission Delegated Regulation (EU) 2015/35, Article 19: Data Used in the Calculation of Technical Provisions. EIOPA Single Rulebook. Article 19.

  • European Union. (2015, as amended). Commission Delegated Regulation (EU) 2015/35, Article 21: Appropriate Use of Approximations to Calculate the Best Estimate. EIOPA Single Rulebook. Article 21.

  • 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. https://doi.org/10.1257/aer.96.4.938. Journal record.

  • IFRS Foundation. (2023). IFRS 17 Insurance Contracts. Effective for annual reporting periods beginning January 1, 2023. Official standard.

  • Long-Term Care Valuation Work Group, American Academy of Actuaries, and Society of Actuaries Research Institute. (2021). Long-Term Care Insurance Mortality and Lapse Study. Study and materials.

  • National Association of Insurance Commissioners. (2017). Actuarial Guideline LI: The Application of Asset Adequacy Testing to Long-Term Care Insurance Reserves. October 5; regulatory guidance and actuarial opinion materials. Guideline text.

  • National Association of Insurance Commissioners. (2026). Long-Term Care Insurance. Insurance Topics, updated July 23. NAIC topic page.

  • National Association of Insurance Commissioners. (n.d.). Long-Term Care Insurance Multistate Rate Review: MSA Framework. Regulatory framework.

  • National Association of Insurance Commissioners. (n.d.). Model Laws: Official NAIC Model-Law Register. Model-law register.

  • Organisation for Economic Co-operation and Development. (2024). Is Care Affordable for Older People? OECD Health Policy Studies. OECD Publishing, Paris. https://doi.org/10.1787/450ea778-en. Research report.

  • Organisation for Economic Co-operation and Development. (2024). Cost-sharing and the Current Adequacy of Long-Term Care Systems. Chapter 3 of Is Care Affordable for Older People? Report chapter.

  • Society of Actuaries. (2015). 2000–2011 Long-Term Care Experience Basic Table Development. Study and tables.

  • Society of Actuaries and LIMRA. (2015). Long-Term Care Intercompany Experience Study—Policy Terminations Aggregate Databases 2000–2011 Report. Study and data.

  • Society of Actuaries, Long-Term Care Experience Committee. (2020). Long-Term Care Intercompany Experience Study—Aggregate Database 2000–2016 Report. August 12. Report PDF.

  • Society of Actuaries and LIMRA. (2021). 2000–16 Individual Long-Term Care Policy Persistency—Lapse and Mortality Rates Report. Study and data.

  • Society of Actuaries Research Institute and LIMRA. (2025). 2000–2023 Long-Term Care Data Request. Research solicitation.

  • Society of Actuaries, LIMRA, and National Association of Insurance Commissioners. (2025, August 4). Industry Leaders Launch New Long-Term Care Insurance Industry Experience Study. Announcement.