Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Long-Term Care Insurance Interest Rate and Asset-Liability Risk

Abstract. Long-term care insurance (LTCI) converts uncertain future dependency, care utilization, and benefit duration into legally specified payment obligations that may remain outstanding for decades. The financial consequence is a compound liability: claim emergence and termination are uncertain, policyholder behavior responds to contractual options and premium changes, benefits may escalate with an inflation rider, and the assets accumulated from early premiums mature or must be reinvested before later benefits are paid. Interest-rate risk therefore enters LTCI through the present value of expected obligations, investment earnings, the pricing of new business, the reinvestment of coupons and maturities, the timing of capital strain, and the affordability of premiums. This monograph investigates those channels as distinct but interacting mechanisms. The analysis combines the formal cash-flow provisions in U.S. actuarial practice and NAIC long-term-care asset-adequacy guidance, international supervisory rules, IFRS 17 measurement, published LTCI experience research, and the economics of private insurance demand. It specifies a policy-state model, derives liability present value and key-rate sensitivities, demonstrates an illustrative duration-mismatch and reinvestment exercise, and distinguishes accounting valuation from economic asset adequacy and solvency. It shows that duration matching is a necessary local diagnostic but an incomplete test of long-horizon solvency when future cash flows are state contingent, premiums and claims are linked, and investment returns incorporate credit and liquidity risk. The decisive practical object is the distribution of available asset cash flows against contractually and behaviorally plausible liability cash flows under coherent joint economic and care scenarios. The investigation supplies a transparent analytical synthesis with reproducible hypothetical calculations. It does not estimate an individual insurer's reserves, establish a universal LTCI interest-rate sensitivity, present proprietary experience as public data, or introduce a new actuarial theory.

Keywords: Long-term care insurance; interest rate risk; asset-liability management; reinvestment risk; key-rate duration; cash-flow testing; asset adequacy; morbidity; lapse; claim termination; care-cost inflation; insurance solvency; IFRS 17; Solvency II; NAIC Actuarial Guideline LI.

Introduction

A promise to pay for long-term care can be issued decades before the insured person begins receiving benefits. Premium income may be collected during working or early retirement years, while the largest expected claim payments may occur at very advanced ages. For traditional stand-alone, level-premium LTCI, the insurer holds assets before claim incidence becomes substantial. The margin between premium accumulation and later care expenditure depends on the term structure of interest rates, the timing and reinvestment of asset cash flows, and the accuracy of actuarial assumptions about morbidity, mortality, lapse, care setting, and benefit escalation. The NAIC long-term-care overview describes the legacy market's combination of challenging experience, repricing, and concentrated participation. Its observations are market-specific, not a basis for extrapolating one universal global LTCI contract.

The problem has a long actuarial history, yet it cannot be reduced to an unfavorable investment yield. Lower interest rates increase the present value of future payments under a current-rate valuation; they can also raise the market value of fixed-rate assets. Whether economic equity rises or falls depends on relative sensitivities, cash-flow timing, hedge structure, credit spreads, and the premium and benefit terms. A higher yield may improve future reinvestment income while reducing the market value of bonds already owned and provoking surrender, liquidity or credit-risk effects. The terms of each valuation framework must therefore be stated before a gain or loss is attributed to a rate movement. IFRS 17's measurement requirements and Solvency II Article 77 illustrate two distinct legally defined calculations; neither can simply be equated with an unconstrained forecast of investable return.

The research question is: How do interest-rate changes and asset maturity structures affect the long-term capacity of an LTC insurer to meet stochastic, often inflation-linked benefits, and how can the effect be identified under real contractual, accounting, and supervisory constraints? The inquiry proceeds by distinguishing three primary objects. A liability projection describes the amount and timing of promised net payments under specified demographic and contract assumptions. A valuation convention assigns a present amount to those projected payments. An asset-adequacy analysis asks whether available invested cash flows and legally attainable future revenues can satisfy obligations across adverse but plausible paths. These objects interact through the same underlying policy block; their numerical outputs answer different questions.

The evidence hierarchy gives priority to original supervisory requirements, actuarial standards, formal reporting standards, published experience-study methods, and peer-reviewed economic evidence. Historical papers explain demand and selection mechanisms; they are not presented as fresh evidence of 2026 interest rates. The Treasury interest-rate statistics provide an observable U.S.-dollar market reference, while their par yields are neither LTCI liability discount rates by definition nor an insurer's realizable portfolio return after expenses, defaults and transaction costs.

The unit of analysis is an issued contract or homogeneous cohort with specified product form, issue age, underwriting, benefit type, coverage currency, premium provisions, inflation rider, elimination period, maximum benefit, and observation date. The analysis preserves distinctions between an eligible care event, a reported claim, a payment, a return to a less-intensive care state, a death, and policy termination. These distinctions define the stochastic flows to which the yield curve is applied.

Conceptual Foundations and Historical Development

Insurance obligations, care needs, and covered events

Long-term care is assistance associated with sustained limitations in daily functioning or cognitive impairment. It can be provided at home, in the community, or in residential institutions. Private LTCI specifies a contractual trigger, waiting or elimination period, reimbursement or indemnity design, benefit maximum, benefit period, and potentially a compound-inflation benefit option. Actual care costs, insured benefits, and insurer payments coincide only under particular contract conditions. A household receiving informal care might have considerable care needs but no reimbursable claim; a claimant might exhaust a lifetime maximum while care continues; an inflation rider might grow a contractual benefit ceiling without matching every local care-service price change. The NAIC Model Regulation makes benefit, eligibility, inflation protection, and reserve design part of the contractual-regulatory setting, although adoption and amendments vary by U.S. state.

For an insurer, the future liability is the contractual benefit cash flow plus allocated claim and administrative costs, reduced by future allowable premium income when appropriate to the valuation basis, with explicit treatment of reinsurance. Statutory reserves, financial-reporting fulfilment cash flows, and management's prospective funding measure use distinct boundaries and allowable assumptions. An actuarial projection that presumes all possible future rate increases will be approved would import an uncertain regulatory decision into the funding base. NAIC Actuarial Guideline LI requires support for future premium increases in LTC asset-adequacy work and a documented management plan with jurisdiction-specific expected approvals and implementation timelines.

The central operational separation is between dependency states, claim states, contract states, and financial states. Dependency measures care need; claims encode eligibility, notification, and payment; contract states encode active premium, waiver, exhaustion, or lapse; financial states describe prices and asset cash flows. The insurer's payment process results from mapping across these dimensions. An insured individual can remain in need of care after a contract terminates, and a contract can remain active while no claim is payable.

The economics of deferred insurance promises

Traditional LTCI premiums collected early in a policy's lifetime finance risks realized much later. Product cohorts are selected at issue; claims are concentrated in older ages; underwriting and underwriting-year mix affect expected future claim timing. High persistency can enlarge the volume of late claims relative to an original assumption that more contracts would terminate without benefit. The SOA's 2021 mortality and lapse study is directly relevant because both survival and voluntary lapse determine which policies remain exposed when later benefits become payable. Mortality and lapse are separate competing exits, not interchangeable reductions in claim incidence.

This delayed financing is particularly sensitive to an assumption that insurers can repeatedly obtain the same portfolio yield. A bond portfolio purchased at issue locks in some contractual cash flows; it rarely locks in the yield on every coupon and maturity that must later be reinvested. In addition, taxable and statutory measures of investment income need not coincide with market-value change. A long-dated care promise may be economically exposed even when current statutory earnings look stable, because the future maturity of assets precedes the tail of claims.

Redington's (1952) life-office immunization analysis established the relevance of first- and second-order valuation sensitivity in historical actuarial finance. Its deterministic immunization conditions motivate a local interest-rate diagnostic, while stochastic care cash flows introduce additional dimensions of risk.

The classical financial distinction between price risk and reinvestment risk applies with exceptional force to LTCI. Falling interest rates generally raise market values of already-owned fixed-rate assets but lower future rates available on reinvested proceeds. Rising rates generally create the reverse initial price movement, subject to credit and optionality. The scale of both effects depends on the portfolio, and their timing matters for solvency. Asset-liability management must simultaneously address economic value, periodic liquidity and admissibility of assets under relevant regulation.

Legacy experience and the discipline of explanation

The market experience of long-standing U.S. private LTCI products provides a caution about assuming independent demographic, behavioral and financial parameters. The NAIC's 2026 LTCI background reports major market adjustments and the need for improved pricing. The Society of Actuaries' LTC experience-study directory identifies historical work on persistency, incidence and mortality, while a 2025 joint industry study announcement explicitly describes planned analysis of policies observed through 2023. An announced study is evidence of a collection program, not evidence that its final experience tables were already published or validated.

Economic research further complicates simplistic experience interpretation. Finkelstein and McGarry (2006), in their study of private information, showed that insurance purchase may reflect both privately perceived care risk and insurance preferences; these channels can offset in aggregate correlations. Brown and Finkelstein (2007), in their market-size analysis, documented the importance of both contract supply and demand in a historically specific U.S. market. These studies support attention to heterogeneous selection and demand rather than a fixed claim that adverse selection always raises every insurer's future claims.

Interest rates are part of the legacy history because old premium assumptions encoded returns expected at issue, and the subsequent yields available to sustain claims could differ materially from that projection. Yet an observed funding deficiency is the joint result of earlier valuation assumptions, later asset choices, realized morbidity, lapse, expenses and approvals. Attribution of a premium increase to interest rates alone requires a counterfactual holding the other relevant determinants constant; otherwise the phrase describes a contributing channel rather than a causal decomposition.

International institutional scope

Private traditional LTCI in the United States occupies a particular place within wider arrangements for financing long-term care. Other countries use different combinations of public social insurance, taxation, means-tested benefits, family support and private cover. It would be invalid to apply U.S. premium-rate-filing rules to a public long-term-care program or to treat all national programs as private insurer liabilities. This monograph therefore uses U.S. statutory and rate-review texts as jurisdiction-specific examples and IFRS 17, the International Association of Insurance Supervisors (IAIS), and EU Solvency II as comparative frameworks for relevant entities. The IAIS Insurance Core Principles address the supervisory linkage between investments, liabilities, enterprise risk management and capital, with implementation depending on jurisdiction. They provide a cross-border discipline for assessing risk categories, rather than a single worldwide reserve formula.

The scope is the financing of insured promises. Demographic changes in public long-term-care spending affect demand and the background cost environment but do not automatically become liabilities of a private insurance contract. The actuarial object is always a defined promise under a defined governing law.

Care Liability Cash-Flow Duration

A policy-state specification

Define time tt in years after valuation. For each contract ii, let Xi(t)X_i(t) denote its joint state, including policy in force, functional-care status, paid-claim status and death. Let bi(t,X)b_i(t,X) denote contractual benefit payments after limits and waiting periods, ei(t,X)e_i(t,X) attributable expenses, pi(t,X)p_i(t,X) collectible premiums, and ri(t,X)r_i(t,X) eligible reinsurance recoveries. The net liability cash flow for a cohort is

CL(t)=∑i{bi(t,Xi(t))+ei(t,Xi(t))−pi(t,Xi(t))−ri(t,Xi(t))}.C_L(t)=\sum_i\{b_i(t,X_i(t))+e_i(t,X_i(t))-p_i(t,X_i(t))-r_i(t,X_i(t))\}.

The expression is a definition of the projected net cash-flow object, not a claim that its stochastic terms can be measured without error. Depending on the reporting basis, reinsurance may be measured as a separate asset and premiums may be constrained by the contract boundary. IFRS 17 explicitly requires consistent identification of contract groups, relevant cash flows, and financial versus nonfinancial risk.

For transparent modeling, use a state set such as active without care; eligible receiving home care; eligible receiving institutional care; temporarily recovered or reduced intensity; benefit exhausted; lapse or surrendered; and dead. A time-inhomogeneous transition matrix P(s,t)P(s,t) or transition intensity generator Q(t)Q(t) specifies conditional probabilities of moving among states. Claim adjudication, waiting periods, benefit exhaustion and premium waiver require additional contract variables or an expanded state space. A semi-Markov approach can represent dependence on time already spent in care, which matters when recovery or termination intensity varies with elapsed episode duration.

The cohort's expected net cash flow is mt=E[CL(t)∣F0]m_t=E[C_L(t)\mid\mathcal F_0], where F0\mathcal F_0 denotes information available at valuation. A simple discounted best-estimate financial component is L0=∑t=1TmtD(0,t)L_0=\sum_{t=1}^{T} m_t D(0,t), with D(0,t)D(0,t) a valuation discount factor for the specified currency, curve and convention. For stochastic financial factors a coherent projection may instead use E[∑tCL(t)D(0,t)]E[\sum_t C_L(t)D(0,t)] with appropriate joint dependence. Financial reporting or statutory methods may add separate risk adjustment, margins, or other quantities; the displayed sum is not a universal definition of a booked reserve.

Incidence, termination and the long tail

Suppose aua_u is the probability per exposure unit of first qualifying claim at time uu, conditional on being eligible and in force; Sc(k∣u,z)S_c(k\mid u,z) is the probability a claim remains payable kk periods later given entry characteristics zz; and Bu,kB_{u,k} is the contractually limited expected payment at that duration. Then the gross expected paid benefits at time tt can be expressed schematically as

E[Bt]=∑u≤tNuauSc(t−u∣u,z)E[Bu,t−u∣payable],E[B_t]=\sum_{u\leq t} N_u a_u S_c(t-u\mid u,z) E[B_{u,t-u}\mid\text{payable}],

with sums across care settings and insured strata implicit. The formula shows why an interest-rate assessment needs both incidence and termination experience. Two cohorts with equal expected lifetime claim totals can have different present values and different asset-maturity needs because their timing differs. A small shift in claim entry to later ages may lengthen the tail; faster care transition or extended survival may move payments differently. Its direction cannot be inferred from a single average age at claim.

Published SOA LTC experience resources provide defined historical observations and methodological caveats. They are not automatically transferable across insurer underwriting, issue decades, benefit inflation riders, care delivery structures, and calendar periods. The experience used must be matched to the risk set and to the policy wording under analysis. An observed claim termination can mean recovery, death, benefit exhaustion or administrative termination, each with different implications for projected benefits.

Timing measures and their interpretation

For nonnegative gross benefit cash flows, define a discounted benefit-weighted Macaulay duration

DL=∑t=1Tt mtD(0,t)∑t=1TmtD(0,t).D_L=\frac{\sum_{t=1}^{T}t\,m_tD(0,t)}{\sum_{t=1}^{T}m_tD(0,t)}.

This duration is a first-moment summary of discounted payment timing. If net cash flows contain negative premiums, the same arithmetic may be unstable or misleading because opposite signs can cause cancellation. Separate durations for gross benefits and future premiums should therefore be computed before discussing net sensitivity. Macaulay duration also differs from effective duration, which includes the response of projected cash flows and contractual options to a yield-curve change.

For a flat annual effective rate yy and fixed positive cash flows, L(y)=∑tmt(1+y)−tL(y)=\sum_t m_t(1+y)^{-t}. Its derivative is dL/dy=−∑tt mt(1+y)−t−1dL/dy=-\sum_t t\,m_t(1+y)^{-t-1}. The modified-duration approximation is ΔL/L≈−DmodΔy\Delta L/L\approx-D_{\mathrm{mod}}\Delta y, where Dmod=DL/(1+y)D_{\mathrm{mod}}=D_L/(1+y). In a multi-rate setting, a single-duration scalar misses curve reshaping; key-rate measures or a vector of sensitivities better represent maturity-specific valuation effects.

A care contract's cash-flow duration is conditional on morbidity, inflation, lapse and premium assumptions. A quoted term to maximum benefit maturity is a contractual boundary, not its expected duration. A twenty-year remaining contract horizon can have a low effective duration if benefits concentrate early, while a lifetime benefit promise can have a much longer and scenario-dependent tail. The correct inference requires the projected payment distribution itself.

Benefit inflation and nominal cash flows

LTC benefit inflation is more complex than multiplying all claims by a general consumer-price index. A policy may provide specified compound growth in a daily reimbursement maximum, a fixed indemnity, or no automatic benefit increases. Actual provider-cost growth can diverge from the benefit cap. The insurer's payable amount may be Bt=min⁡{Kt,Ct}B_t=\min\{K_t,C_t\} for contractual ceiling KtK_t and eligible care cost CtC_t, subject to limits. The expectation of a capped payment is not generally the smaller of expected cost and expected cap. Changes in care setting and utilization can also alter nominal amounts independently of price inflation.

Nominal liability discounting requires nominal benefit projections. Real discount rates may be used with consistently deflated cash flows and an explicitly modeled price index. Applying nominal care escalation to cash flows and a real discount rate without conversion double-counts inflation. Similarly, a nominal Treasury rate cannot be mechanically subtracted from a care-cost growth estimate to produce a contract-specific real yield; basis, currency and uncertainty must first be aligned.

The methodological result is that liability duration has to be calculated after modeling contractual cap mechanics and the timing of claim states. It cannot be read directly from demographic longevity, nor from a generic fixed-income benchmark.

Discount Rates and Reinvestment Scenarios

The meaning of an interest-rate assumption

An interest rate enters LTCI work in at least four different capacities. A contractual accumulation or crediting rate governs a specified product feature where one exists. An asset yield is the expected or realized return associated with invested resources. A liability discount rate converts future estimated payments into a valuation amount under a defined measurement regime. A stress-rate path tests the cash-flow consequences of future financial states. The four can differ while remaining internally consistent. Confusing an original issue-date pricing yield with a current financial-reporting discount curve is a particularly consequential error for legacy contracts.

In U.S. asset adequacy work, NAIC Actuarial Guideline LI distinguishes cash-flow testing with an explicit allocation of investment income from gross-premium valuation approaches that do not model assets. Under the latter, the discount rate must take account of yields on current assets supporting the liability and on future assets bought with premium income and reinvestment proceeds. This requirement has a precise financial implication: using a flat historical portfolio yield to value every future cash flow can misrepresent late-period earnings when the underlying holdings mature substantially earlier than the care claims.

For IFRS reporting, AASB 17 paragraph 36, an accessible adoption of IFRS 17's discounting requirements, specifies time value, cash-flow and liquidity characteristics, consistency with observable market prices, and exclusion of factors affecting traded instruments but not insurance contracts. The IFRS Foundation's IFRS 17 summary also distinguishes fulfilment cash flows, nonfinancial risk adjustment, and contractual service margin. Current IFRS measurement must therefore be distinguished from projection of the expected book yield of an existing asset portfolio.

Under the European Solvency II regime, Article 77's best estimate uses probability-weighted future cash flows and the relevant risk-free term structure; technical provisions normally combine that best estimate with a risk margin. Article 77e's published curves formalize official technical information, including applicable adjustments. A matching adjustment can be used only where the legal portfolio and supervisory conditions are met, as stated in Article 77b. An LTC insurer cannot infer matching-adjustment eligibility solely from owning a bond portfolio. Differences between the U.S., IFRS and EU measures arise partly because their objectives and permitted rates are different.

A reproducible zero-coupon duration mismatch

Consider a stylized insurer with one fixed liability payment of $10 million exactly twenty years from valuation. Assume annual effective discounting at a constant 3% and no claims uncertainty, premium cash flows, expenses, default, tax or contract options. This deliberately stripped-down example isolates the financial maturity effect. The liability present value is

L0=10,000,0001.0320.L_0=\frac{10{,}000{,}000}{1.03^{20}}.

Fund it initially with a ten-year zero-coupon asset whose present value exactly equals L0L_0. The asset's face value is the L0L_0 amount compounded for ten years at 3%, equivalently $10{,}000{,}000/(1.03)^{10}$. Both initial values agree by construction, but the asset matures ten years earlier than the single liability. The market-value surplus is zero initially, while exposure to future reinvestment and relative duration remains.

At a parallel instantaneous rate shock to 2%, the market value of the liability rises by more than the market value of the shorter asset. At 4%, the reverse occurs at the immediate valuation date. These are mark-to-market calculations with a fixed bond cash flow; they are not proof of what will happen after a sequence of future rates, asset sales, defaults or care claims. The comparison is correct only while the hypothetical liability is fixed and the discount rate applied consistently to both legs.

The derivative of surplus S(y)=A(y)−L(y)S(y)=A(y)-L(y) is dS/dy=−DAmodA+DLmodLdS/dy=-D_A^{\mathrm{mod}}A+D_L^{\mathrm{mod}}L. If A=LA=L at the starting yield, the first-order change becomes ΔS≈L(DLmod−DAmod)Δy\Delta S\approx L(D_L^{\mathrm{mod}}-D_A^{\mathrm{mod}})\Delta y. A longer liability than asset consequently produces a negative surplus movement from falling yields. The conclusion depends on the relative durations and values, not the descriptive name of the product. It reverses when the asset side is more rate sensitive, and it changes when credit spreads or contingent insurance payments move with rates.

Reinvestment is a path-dependent risk

Take a distinct hypothetical asset that can earn 3% for ten years. Suppose the planning assumption was that its accumulated value would then continue earning 3% for another ten years, whereas a subsequent reinvestment environment permits only 1%. The planned twenty-year accumulation factor is (1.03)20(1.03)^{20}; the realized factor under this simplified two-period path is (1.03)10(1.01)10(1.03)^{10}(1.01)^{10}. Their ratio is (1.01/1.03)10(1.01/1.03)^{10}. A reduction in terminal proceeds arises even without any mark-to-market sale loss and even if the original ten-year holding delivered every promised payment. This is the essence of reinvestment risk.

In a real LTC portfolio, assets often provide coupons, principal repayments, callable obligations, mortgages, structured cash flows and receivables rather than one maturity. Floating-rate assets reprice with benchmarks; callable bonds can shorten in falling-rate environments; prepayments and policyholder flows can change the cash available for reinvestment. Market yield and realizable asset return diverge after defaults, downgrades, expenses, illiquidity and reinvestment frictions. Matching asset coupon cash flows to the expected mean benefit stream leaves a residual mismatch when claims arrive earlier or later than the mean.

The Treasury yield-curve publication supplies a reproducible observable curve input for U.S.-dollar illustration. The par curve is not a zero-coupon discount curve, however, and should not be treated as one without bootstrapping or an appropriate term-structure conversion. An insurer needs currency-specific, valuation-date-specific assumptions and should test the sensitivity to tenor interpolation and extrapolation, especially when contract claims extend beyond liquid bond maturities.

Nonparallel curves, basis risk, and optionality

A parallel one-percentage-point rate change is a useful teaching experiment; realistic shocks can steepen, flatten or twist the curve. Define the value of a liability under key rates yky_k as L(y)L(\boldsymbol y). A first-order change is ΔL≈∑k(∂L/∂yk)Δyk\Delta L\approx\sum_k(\partial L/\partial y_k)\Delta y_k. Replacing the entire key-rate vector with a single duration number suppresses whether the mismatch is concentrated at the ten-year, twenty-year or longer maturities. A liability with a forty-year tail may be particularly sensitive to extrapolation assumptions rather than to any one traded security.

Credit spread is an additional dimension. A corporate bond can gain market value when risk-free rates decline while losing value if its credit spread widens. LTCI claims are generally not reduced merely because the insurer's credit spread widens. A liability discount methodology that imports an asset spread without a contractually justified liquidity or risk adjustment would falsely reduce reported liabilities when the insurer becomes riskier. This is one reason current insurance reporting rates and asset portfolio yields require conceptual separation. In Solvency II, the matching-adjustment restrictions and volatility-adjustment provisions are legally specified and differ by use case; they cannot be added indiscriminately to the same obligations.

Optionality enters through assets, liabilities and regulatory actions. An issuer may call a high-coupon bond after rates fall, creating reinvestment demand at precisely the least favorable time. Policyholders may retain valuable LTCI coverage despite premium changes, with lapse behavior varying by age, accumulated benefits and ability to replace coverage. An insurer may seek premium increases, but those increases can be delayed, reduced or conditioned on benefit alternatives. NAIC's multistate rate-review framework supplies a coordination mechanism for certain U.S. rate requests, not a guarantee of approval or a source of unrestricted future revenue.

Joint financial and care-state scenarios

A defensible stress model should vary the rate curve jointly with the variables that can transmit a common macroeconomic environment into care claims and asset values. Example scenario families include prolonged low rates with heavy late claims; high nominal rates with elevated care-cost inflation; credit-spread widening with constrained asset liquidity; adverse claim persistence with delayed premium relief; and asymmetric yield curves with insurer bonds maturing before the claim tail. Each path must specify a measurement date, starting balance sheet, asset universe, contract design and investment policy. A scenario that changes benefit inflation while keeping a fixed nominal discount curve tests one dependency but does not represent every plausible joint inflation-rate relationship.

Dependence should be modeled cautiously. Long-term-care incidence is not mechanically a function of nominal bond rates. Correlation arising through economic conditions, wealth effects, care prices, policyholder behavior or insurer investment responses is empirical and time-varying. Treating it as a universal deterministic relationship would create false precision. The ASOP No. 56 modeling guidance calls for assumptions appropriate to purpose, transparent uncertainty, and consistency of significant assumptions.

Scenario paths should therefore report both the estimated direct valuation movement with care cash flows held constant and the additional changes from behavioral or demographic assumptions. This decomposition is a diagnostic, not a validated causal attribution. It makes clear where empirical evidence is needed and where results depend on a scenario chosen for supervisory prudence.

Asset Adequacy and Solvency

Cash-flow sufficiency is a horizon-wide question

At any valuation date the insurer's assets can exceed a prescribed reserve and still be exposed to insolvency or costly asset liquidation under adverse cash-flow timing. The relevant funding question is whether admissible asset receipts, realistic future premiums and other available resources can meet claims, expenses and capital needs across scenarios without relying on unavailable future revenue. For a scenario ω\omega and horizon tt, let At(ω)A_t(\omega) be usable market- or regulation-consistent asset resources, Pt(ω)P_t(\omega) premium receipts, Bt(ω)B_t(\omega) benefits, Et(ω)E_t(\omega) expenses, and Rt(ω)R_t(\omega) reinsurance settlements. A stylized liquidity recursion is

Ht+1(ω)=Ht(ω)+It(ω)+Pt(ω)+Rt(ω)−Bt(ω)−Et(ω)−Kt(ω),H_{t+1}(\omega)=H_t(\omega)+I_t(\omega)+P_t(\omega)+R_t(\omega)-B_t(\omega)-E_t(\omega)-K_t(\omega),

where ItI_t denotes net cash from invested assets or reinvestment operations and KtK_t captures other explicitly modeled capital or financing requirements. The recursion is a cash-flow accounting device; its precise definitions must prevent counting proceeds both in HtH_t and again in ItI_t. Market-value surplus and readily spendable cash are related but distinct quantities.

An asset-adequacy test should identify scenarios in which HtH_t breaches a minimum operating or statutory constraint, not merely those with negative discounted terminal values. Selling bonds to pay early care claims can crystallize price or credit-spread losses. Conversely, an insurer may have enough present-value assets but inadequate near-term liquidity. Claims settlement patterns, collateral calls, reinsurance payment lags and operational constraints all matter. A reserve margin is not an unlimited substitute for cash.

The revised ASOP No. 7 became effective for relevant actuarial work products issued on or after June 1, 2026, and addresses the scope, methods and circumstances of life or health cash-flow analysis. Its distinction between analyzing assets, liabilities or both is central. ASOP No. 22 addresses opinions based on asset adequacy and discusses cash-flow testing and alternative methods. These standards govern specified actuarial services in the United States; the monograph uses them to understand methods, not as universal legal requirements.

U.S. LTC-specific asset-adequacy treatment

Actuarial Guideline LI is unusually explicit about LTCI. It applies, under the identified model framework, to companies exceeding 10,000 in-force lives in the affected LTC contract scope; combination products for which substantial risk is life insurance or annuity risk may be outside its stated scope. Its text requires LTC-specific analysis, addresses the conditions under which deficiencies may be aggregated against other blocks, and prescribes documentation of mortality, lapse, morbidity, interest-rate and anticipated premium-increase assumptions. Regulatory implementation and subsequent valuation-manual changes must be checked at the actual state and valuation date.

This requirement provides a direct answer to the central research question. An insurer cannot demonstrate asset adequacy solely by projecting gross benefits using a single favorable yield. When an LTC block's material asset-related risks warrant cash-flow testing, the model must represent the actual economic resources and their allocation to the block. If a simplified valuation is used, its discount rate must nevertheless reflect both current supporting assets and future reinvestment or purchases. Assumed rate increases require documented management and jurisdictional support; they are not fungible with contractual premiums already owed.

The conceptual difference between stand-alone testing and companywide aggregation matters for multi-line insurers. A profitable annuity or life block may provide a source of group strength, but legal separation, fungibility, investment restrictions, timing and method compatibility determine whether its resources can credibly offset an LTC deficiency. Supervisory permissions cannot be inferred from consolidated economic surplus alone. The asset adequacy memorandum should disclose the precise method and any allowed offset.

Accounting, statutory solvency and economic equity

Under IFRS 17, the current estimate of fulfilment cash flows is combined with nonfinancial risk adjustment and, where applicable, contractual service margin. Financial assumptions and discount rates are updated under the standard's reporting requirements, while insurance service results and finance results receive distinct presentation. An asset-liability movement can therefore affect reported earnings, other comprehensive income and equity differently depending on applicable accounting choices and asset classification. The IFRS liability is a measurement of insurance contracts, not a stand-alone measure of whether an insurer holds enough liquid assets to satisfy every stressed claim.

Under EU Solvency II Article 77, technical provisions ordinarily comprise best estimate and risk margin; valuation also supports supervisory capital analysis. EIOPA's July 2026 guidance update discusses a revised time-dependent risk-margin calculation, with the amended guidelines scheduled to apply on January 30, 2027. Therefore a 2026 valuation should not be described as already governed by those future application dates merely because the guidance was published earlier. The outcome of any EU LTC portfolio assessment also depends on its contract classification and the applicable solvency rules.

IAIS ICP 15 on investments emphasizes the interdependence of insurer assets and liabilities and qualitative as well as quantitative investment controls. The wider ICP framework connects valuation, enterprise risk management and capital adequacy. A management dashboard that reports only a single economic duration gap would therefore omit credit concentration, liquidity, counterparty, operational and basis risks that may jointly drive insurer failure.

A disciplined comparison uses four parallel balances: contractual cash-flow projection; reporting-standard insurance liability; local regulatory technical reserves and admissible capital; and economic market-value surplus with liquidity paths. These measures should be reconciled through identifiable assumption and rule differences. Neither the highest nor lowest number is automatically the correct measure for every decision. Board decisions concerning hedging, reserves, capital or policyholder protection require the appropriate object.

Capital risk and reverse stress testing

A forward scenario starts with a specified rate, credit and claims path and projects the balance sheet. A reverse stress test begins with failure conditions—reserve deficiency, cash shortage, capital threshold, or breach of collateral terms—and searches for plausible joint paths that cause the breach. Reverse testing can reveal that apparently moderate interest-rate stress becomes material only after a lapse shortfall, increased duration of claim payments, or investment downgrades. The exercise should also model management actions and their legal feasibility; treating a future action as guaranteed makes the scenario circular.

An LTC portfolio with fixed nominal benefits but long payment timing is highly sensitive to rates through valuation. A portfolio with robust inflation-linked benefits has additional real-resource exposure; its risk cannot be hedged by a nominal-duration match alone. A portfolio funded by illiquid credit may have high projected spread income and simultaneously be exposed to defaults, impaired secondary markets and forced sales. More expected return is not interchangeable with more safe capacity to pay policyholders.

Capital allocation should recognize correlations among LTCI and other insurer risks. In stressed macroeconomic environments, widening credit spreads and adverse care-cost inflation may coincide. Alternatively, these movements can be weakly related. Model governance should separate empirical dependence estimates, regulatory standard-formula assumptions and deliberately conservative scenarios. Any claimed diversification benefit must be supported by its measurement basis, and any capital figure must identify the applicable jurisdiction and calibration.

Rate-review governance and the policyholder

Future premium increases alter both insurer assets and policyholder outcomes. The U.S. multistate review process was designed to improve coordination among states, but the final decision and timing remain subject to applicable regulatory procedures. The same rate increase can trigger different lapses or reduced-benefit elections across age and wealth groups. A model that books the entire anticipated premium uplift while retaining all insureds on unchanged benefit terms will generally be internally inconsistent unless there is specific evidence for that response.

For legacy business, the timing of rate approval creates financial optionality. A request might improve cash receipts only after months or years; approvals may be phased, smaller than requested, or accompanied by reduced-benefit choices. An insurer that relies on those receipts to meet immediate claim payments is exposed to implementation risk. The policyholder simultaneously bears affordability and coverage risks. Long-term care asset-liability management therefore has a direct consumer-protection dimension beyond the insurer's reported surplus.

This observation also explains why actuarial stress tests should distinguish a regulatory rate decision from a projected policyholder response. An approved increase is an observable legal event, whereas the consequent lapse or benefit-election behavior must be estimated. The two form a sequence whose delays and feedback effects can materially influence the asset-liquidation path.

Cross-jurisdiction interpretation

The United States, the EU reporting framework and IFRS-based accounting are useful for distinct reasons. The U.S. provides a specific long-duration private LTCI supervisory case; EU Solvency II supplies an explicit market-consistent technical-provision structure; IFRS 17 supplies a general insurance-contract accounting measurement framework. One jurisdiction may use more than one of these frameworks for different purposes. Cross-jurisdiction conclusions should identify the entity type, contract class, legal adoption, reporting date, discounting convention and capital definition before comparing liabilities or apparent sensitivity.

The resulting synthesis is portable at the level of cash-flow logic. The exact regulatory numbers are not portable without reconstruction. A private LTC product issued under U.S. state regulation cannot be assumed to have the same prescribed reserve or permitted liability adjustment as a European life or health insurer, and neither can be equated with a tax-financed public care program.

Integrated Findings and Research Agenda

Interest-rate losses have several transmission paths

The synthesis establishes that an LTCI interest-rate shock can be transmitted through five identifiable routes. The first is valuation: future benefits change present value under a new discount curve even without a change in their nominal amount. The second is investment income: reinvested or newly acquired assets earn rates that differ from issue-date expectations. The third is liquidity: the timing of claims and maturities can force asset sales or borrowing when rates or spreads are unfavorable. The fourth is behavioral and regulatory: premiums, reduced-benefit elections, persistency and approval lags can alter net future cash flows. The fifth is capital and reporting: the same economic event can appear differently under current insurance accounting, prescribed statutory reserving and solvency capital requirements.

This classification is an organizational synthesis of established actuarial and financial mechanisms, not a proposed theoretical system. Each route can be identified by a change in inputs and outcomes within standard cash-flow and balance-sheet analysis. The important substantive conclusion is that the routes cannot be assumed to move in parallel. A falling-risk-free-rate path can benefit mark-to-market assets and harm reinvestment returns; a rising-rate path can improve long-term earnings while creating an immediate asset-value and liquidity strain. Care utilization can aggravate either path through unexpected payment timing.

Matching a mean duration cannot guarantee claim funding

Duration matches the first-order sensitivity of values to a narrowly defined rate movement; it does not match the distribution of liability payments. Convexity, curve shape, contractual options, benefit inflation, mortality, claim emergence, credit spreads and future policyholder actions can all break an apparent immunization. A duration-neutral portfolio can remain vulnerable to large yield changes or changes in claim-state probabilities. Exact cash-flow matching is also more demanding than equal durations and may be unattainable for extended, uncertain care obligations.

The asset-adequacy perspective therefore evaluates cash-flow paths and downside funding outcomes in addition to present-value sensitivities. ASOP No. 7 cash-flow guidance and NAIC LTC asset-adequacy guidance provide institutional anchors for this result in U.S. actuarial practice. The general financial reasoning extends to long-duration insurance contracts elsewhere, subject to each jurisdiction's requirements.

Interest-rate and morbidity misestimation must be decomposed

A legacy LTCI insurer that requires rate relief after decades of lower returns and adverse care experience needs a decomposed explanation. The change in required funding should be calculated under alternative interest-rate assumptions holding care experience constant, and under alternative care assumptions holding financial conditions constant, before measuring interaction. This approach distinguishes a model sensitivity from an observed cause. It also shows whether improved future returns could realistically close a legacy funding gap without regulatory rate changes or reduced benefits.

The need for decomposition follows from the source evidence. Experience studies provide mortality, lapse and claims inputs; financial market data describe rates; policy filings and contractual documents define permissible cash-flow revisions. None alone measures the entire balance-sheet deficiency. A credible report reconciles these evidence classes and states what remains unidentified.

Comparative standards lead to a consistent analytical object

U.S. actuarial asset-adequacy testing, IFRS 17 insurance liability measurement, Solvency II technical provisions, and IAIS supervisory risk principles differ in objective and detailed requirements. Their common analytical foundation is the mapping of defined contractual cash flows to valuation and funding conditions. The comparison supports a unified research object—state-contingent insured payment obligations supported by specific asset resources—without asserting a uniform statutory capital requirement across countries.

The supervisor's question and the investment committee's question overlap, but neither is a substitute for the other. A sufficiently funded statutory reserve can coexist with a market-value duration gap, and a market-value hedge can coexist with short-term illiquidity or an accounting mismatch. Both outcomes require attention in governance.

Research questions that can be empirically answered

The next stage of research requires insurer- or cohort-level access to matched policy and investment histories. First, which components of liability duration are robust across underwriting cohorts when care-state transition risks are controlled? Second, how do benefit inflation riders change the covariance of claim amounts and interest rates across inflation regimes? Third, what share of LTC block asset-adequacy deficiency arises from reinvestment rather than claim incidence, termination or premium assumptions, when evaluated under the same model? Fourth, how do approval delays and reduced-benefit elections respond to rate increases by age and contract type? Fifth, can key-rate hedges and longer-maturity assets reduce tail liquidity stress after credit default, collateral and forced-sale effects are included?

These questions admit longitudinal experience studies, stress-based balance-sheet modeling, and credible quasi-experimental work where regulatory and portfolio histories permit. They should be tested with clear denominators, separate identification of financial and care-state shocks, and out-of-sample validation. Their articulation is a research agenda derived from the evidence reviewed; it carries no claim of having obtained answers from inaccessible insurer data.

Principal limits of this monograph

The primary institutional sources establish practice and legal frameworks; they are not evidence of a particular insurer's portfolio profitability. Published experience studies contain historical cohort and design limitations. U.S. private LTCI experience is not representative of all public and private international care-financing systems. The deterministic numeric experiments show mechanisms but omit demographic variance and credit losses; neither is a risk-capital estimate. The current research contains no external independent peer review, insurer-specific internal controls, or causal empirical identification.

These restrictions are part of the substantive result. An apparently simple LTC interest-rate exposure cannot be quantified responsibly without reconciling contract, claims and asset records for the same date and cohort. The methods remain available, and their required evidence has been identified.

Conclusion

Long-term care insurance interest-rate and asset-liability risk is the risk that future contingent care-related payments and available supporting financial resources change in value, amount, accessibility or timing in ways that impair an insurer's capacity to fulfill its contractual promises. The central financial mechanisms are duration and yield-curve exposure, reinvestment after asset maturity, the interaction of credit and liquidity with care payment timing, and conditional premium and claim behavior. Each mechanism is observable in principle, requires its own assumptions, and influences different reporting and solvency measures.

The formal analysis demonstrates why a long-dated care liability funded by shorter assets loses economic value when yields fall and why a reduction in reinvestment rates can impair future accumulation even if the original assets pay in full. The same mathematics shows why equal initial present values do not imply equal exposures. The hypothetical calculations are exact conditional on their specified assumptions; the empirical magnitude for an insurer depends on its actual assets, policy conditions and experience.

The international comparison confirms the need to maintain distinctions among IFRS 17 fulfilment-cash-flow valuation, EU Solvency II best estimate and risk margin, U.S. statutory LTC asset-adequacy requirements, and internal economic funding and liquidity tests. The correct assessment begins with contractual payment states, reproduces the associated cash flows, assigns discounting and investment assumptions consistent with each question, and tests the resulting balance sheet over coherent adverse paths. A source-based scientific account of LTCI asset-liability risk consequently links actuarial experience, financial market structure, contract law and insurer governance in one transparent analytical evaluation.

The original contribution of this monograph is its documented comparative and mathematical synthesis of those established mechanisms, including an explicit separation of valuation, reinvestment, timing, behavioral and regulatory channels. Its conclusions support reproducible risk analysis and define the empirical evidence needed for further work without advancing an unsupported new theory.

References