Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Period-Certain and Premium Refund Annuity Guarantees

Abstract. Period-certain and premium-refund annuity guarantees specify what happens to promised income and unrecovered contractual consideration when a life annuitant dies early. They govern the allocation of payments between the annuitant and beneficiaries, the duration and timing of the insurer's obligations, and the premium required to combine lifetime income with estate-oriented protection. This monograph examines the contractual, actuarial, economic, fiscal, regulatory, and operational dimensions of these guarantees. It distinguishes a life annuity with a fixed minimum payment period from cash-refund and installment-refund arrangements; derives present-value identities under explicit payment and death conventions; constructs reproducible examples; and identifies the effects of mortality selection, term structure, inflation, beneficiary administration, taxation, and financial reporting. The analysis finds that the guarantee is an allocation rule over states of survival and death, not simply an advertised minimum amount. Cash and installment settlement can deliver the same undiscounted total but different economic values and liquidity profiles. A period-certain option shifts early-death exposure toward deterministic payments while retaining longevity exposure thereafter. The study integrates primary regulatory texts with actuarial analysis and the published literature, distinguishes institutional requirements by jurisdiction, and specifies the empirical information needed for further product-specific research. No new proprietary theoretical system is proposed.

Keywords: Life annuity; annuity certain; period-certain guarantee; cash refund; installment refund; annuitant mortality; beneficiary; early-death transfer; mortality credits; premium pricing; actuarial present value; discounting; annuity taxation; IFRS 17; Solvency II; retirement income.

Introduction

Retirement income contracts distribute purchasing power across an uncertain future lifetime. The central life-annuity promise is a stream payable while the covered person survives. A contract can add a minimum number of payments or an obligation to return defined consideration if death occurs before a contractual recovery threshold. The resulting payments may accrue to the annuitant, a named beneficiary, or an estate. Each provision changes both the economic exposure borne by the insurer and the distribution of wealth among individuals and generations. State insurance authorities explicitly distinguish life income with a period certain from life income with a refund. Florida's consumer annuity definitions further separate cash refund, installment refund, and certain-and-life contracts.

The research object is the guaranteed payment provision attached to a life-contingent income contract. Its economic consequences depend on the benefit schedule, meaning of premium, death and payment timing, beneficiary rights, discount curve, interest rates, expense loadings, and mortality. The analysis centers on nonvariable level-payment structures for transparent actuarial derivations, while identifying extensions to increasing, indexed, deferred, joint-life, and pension contexts. A standalone period-only annuity, a pure investment withdrawal plan, and an account surrender right are discussed only to clarify distinct obligations. Product-level conclusions must be tied to an issued contract, applicable jurisdiction, and valuation date.

The method combines verified institutional sources, original published research, explicit state-contingent cash-flow modeling, analytic comparisons, and independently recomputable illustrative figures. The historical and economic foundation of annuitization is traced to Yaari's 1965 lifetime-consumption analysis, with later welfare and selection studies used for the interpretation of guarantees. The analysis of legal and tax consequences uses the exact source jurisdiction rather than extrapolating one state's guidance globally. Empirical claims are restricted to the scope of available studies and summaries. The research cutoff is November 15, 2026.

Scientific Foundations, History, and Research Identity

The contract as a contingent entitlement

A life annuity promises a sequence of payments contingent on the continued life of one or more persons. Under a simple single-life contract, a scheduled payment falls due only when the annuitant is alive at its relevant determination date; contract language determines how payments due but unpaid at death are treated. A period-certain rider modifies this entitlement: for an agreed minimum period, payments are owed regardless of whether the original annuitant survives; later payments remain contingent on survival. A premium-refund rider instead sets a recovery benchmark and pays a defined residual amount after early death. Both features create obligations to someone other than the original income recipient in some states, yet their payment rules differ fundamentally.

The New York insurance guidance states the operational distinction: a life annuity with period certain pays for life and for at least the specified term; a life-with-refund contract compares total benefits already paid against the purchase price and settles the difference, potentially in a lump sum or installments. The Florida insurance description makes the destination of the cash refund explicit. These are authoritative examples of U.S. regulatory consumer explanations, not universal model policy language. Actual contracts can specify net premiums, rounded installments, premium taxes, commutation values, exclusions, and payment dates differently.

A beneficiary guarantee is contractual rather than an assertion that the original annuitant can withdraw the remaining premium whenever desired. For a single-premium immediate annuity, annuitization ordinarily exchanges an accessible asset balance for the insurer's payment obligations. Consumer disclosure material differentiates scheduled income from account withdrawal rights, and the OECD's payout-phase study analyzes the separate economic value of lifetime security and liquidity. Guaranteed nominal payments also do not necessarily maintain purchasing power if inflation raises prices while the contractual sum remains level.

Institutional and historical setting

The economic premise of annuitization is survival-contingent pooling: payments that cease at an annuitant's death help finance benefits to survivors in the pool. In Yaari (1965), optimal consumption under uncertain lifetime demonstrates the relevance of life-contingent markets under stylized assumptions. Davidoff, Brown, and Diamond (2005) examine how optimal annuitization changes when markets are incomplete, bequests matter, and contingent risks remain. A guaranteed residual to beneficiaries can address a bequest or premature-death concern, but it changes the allocation of mortality credits. Its value depends on both household preferences and the insurance price.

OECD (2024) on retirement payout design places lifetime protection within broader retirement-income adequacy, asset liquidity, health status, essential spending, and public pension coverage. This perspective prevents a generic assumption that maximal bequest protection produces maximal retirement welfare. The OECD Recommendation on DC plan design similarly frames the payout phase as part of an integrated retirement system. Period-certain and refund provisions operate at the interface of those goals because they redistribute the premium's economic value across retirement consumption and residual estate protection.

Contract taxonomy and boundary conditions

A pure life annuity terminates future contingent payments after death, subject to any accrued but unpaid amounts. A life annuity with an m-period certain guarantee assures each of the first m payments; after m, income continues only while the annuitant lives. A cash-refund life annuity pays the defined unrecovered guaranteed base at qualifying death in one settlement, according to the contract's death-benefit and processing provisions. An installment-refund life annuity pays the residual in scheduled installments, usually matching the periodic benefit, with a possible reduced final installment. The term-certain-only annuity pays for a fixed period regardless of survival, without continuation for a longer life; its life-contingent element is absent.

The Insurance Compact's immediate annuity standard is particularly important because its scope explicitly includes immediate nonvariable life-contingent contracts with and without period certain, as well as non-life-contingent contracts. This is a uniform product standard for its defined compact filing environment, not a universal life-annuity code. Its record identifies January 8, 2018 as the effective date of the operative 2017 amendment; a further amendment is stated to become effective December 2, 2026 and is future-dated at this dossier's November 15 cutoff. Its specifications provisions require relevant payment commencement, intervals, amounts, and guarantee period information. The standard's beneficiary and commutation provisions demonstrate why precise contract wording matters.

A deferred paid-up income annuity can also include a guarantee period or refund feature. The Compact's deferred income annuity standard explicitly identifies a period-certain annuity and addresses premium-return death benefits. Yet the timing of death before income commencement creates a separate contractual state. The present derivations begin at income commencement and do not silently substitute for a precommencement death-benefit valuation. Likewise, a joint-life continuation clause is a distinct two-life contingency and cannot be replicated merely by extending the single-life certain period.

Period-Certain Guarantee Mechanics

A precise discrete-time cash-flow convention

Let time zero be annuity commencement. Scheduled payments are at the ends of years k = 1, 2, ... ; B is the level annual payment. Let T be the future lifetime measured from commencement, and let S_k = Pr(T ≥ k) be survival to the kth payment date, with S_0 = 1 and a nonincreasing survival sequence. Write v = 1/(1+i) for a positive constant annual effective interest rate i. For a mortality process in which death can occur between payment dates, let q_k* = S_k − S_{k+1} be the probability of death between times k and k+1. Payment and settlement timing assumptions are essential; other conventions produce different present values.

For a pure life annuity-immediate, the expected present-value factor per unit of annual benefit is

alife=∑k=1∞vkSk.a_{\mathrm{life}}=\sum_{k=1}^{\infty}v^k S_k.

The insured cash-flow indicator at time k is 1{T ≥ k}. A level payment B has actuarial present value B a_life, before expenses, profit margins, tax, and any basis or capital adjustments. Under a life-plus-m-certain arrangement the kth payment is unconditional for k ≤ m and survival-contingent for k > m. Its expected present-value factor is

am-certain+life=∑k=1mvk+∑k=m+1∞vkSk.a_{m\text{-certain+life}}=\sum_{k=1}^{m}v^k+\sum_{k=m+1}^{\infty}v^k S_k.

Subtracting the life-only factor yields the transparent cost decomposition

am-certain+life−alife=∑k=1mvk(1−Sk)≥0.a_{m\text{-certain+life}}-a_{\mathrm{life}}=\sum_{k=1}^{m}v^k(1-S_k)\geq 0.

This is a mathematical identity under the stated timing and survival assumptions. The added expected value comes from guaranteed installments that remain payable at early death. It does not require an invented mortality-credit theory. For equal premium and consistent pricing assumptions, a positive extra guarantee value reduces the feasible annual benefit relative to life-only payments. For a fixed annual benefit, the corresponding premium rises. The amount depends on survival, interest, expense and capital assumptions; the inequality alone is not a market quotation.

Event-state analysis and timing

A practical state table must separate (i) alive before the end of the guaranteed term, (ii) dead with guaranteed installments remaining, (iii) alive after the term, and (iv) dead after the term. In the first state B accrues under both the survival and guarantee obligations. In the second state B accrues to the designated contractual recipient for each remaining guaranteed payment date. In the third state B continues as a life payment. In the fourth, future payments cease, absent another surviving-life or independently accrued obligation. The insurer should also record the transition date, the last scheduled payment paid, a death-notification lag, and beneficiary status.

The economic value of an early-death beneficiary stream depends on whether the contract continues each installment at its scheduled time or permits commutation. Optional commutation needs its own discount rate and administrative definition; the cash value of a future series need not equal its nominal total. This distinction is supported by the Compact's commutation provision. Assuming automatic commutation in a contract that allows only installment continuation would alter both the beneficiary's rights and the liability's timing.

Monotonicity and limiting cases

At m = 0, a life-plus-certain arrangement reduces to a pure life annuity under this convention. If m increases by one, its expected value grows by v^(m+1)(1−S_(m+1)). If every covered person survives through the guaranteed period, the incremental actuarial cost is zero at those durations. If survival probability is low, the incremental cost is larger. For a long enough guarantee, most early payments become almost deterministic and longevity-contingent value remains primarily in the tail beyond the guaranteed term. This statement concerns cash-flow sensitivity under a fixed mortality basis, not the premium that a particular insurer will actually quote.

A selected annuitant population with higher survival probabilities reduces the incremental death component for a fixed term, while raising the value of the underlying life annuity. Those effects must be calculated together. Market participants cannot infer a lower total premium solely from one derivative. Finkelstein and Poterba (2004) find relationships between ex post mortality and annuity characteristics in the U.K. market, including estate-related features. Their evidence justifies product-specific selection analysis; it supplies no universal mortality multiplier for every refund contract.

A reproducible five-year demonstration

For illustration only, assume i = 3% per year and S_k = 0.95^k for every integer k, which implies a constant annual survival factor of 0.95. This intentionally simple synthetic mortality schedule is neither a population forecast nor a quotation basis. Geometric summation gives a_life = 0.95/(1.03−0.95) = 11.875000. The five-year guaranteed factor is 12.5059162145 and the ten-year factor 13.8207114300. At a hypothetical net premium of $100,000 with zero expenses and no risk margins, the corresponding annual benefit levels would be $8,421.05 for pure life, $7,996.22 for life-plus-five-certain, and $7,235.52 for life-plus-ten-certain. The monotonic ranking follows from the guarantee value; the monetary figures come only from the synthetic assumptions.

In an issued contract, the guaranteed term may be denominated in months, the first installment may be immediate rather than deferred, fractional periods may apply, and a beneficiary's estate may have a commutation option. A monthly annuity-due implementation requires payment-by-payment recalculation and cannot simply divide the annual amount or present-value factor by twelve. Correct valuation preserves the real schedule, all coverage dates, and the insurer's actual promise.

Cash Versus Installment Refunds

Contractual base, crediting, and trigger

For a typical gross-premium refund structure, let P denote the contractually refundable reference amount at commencement, and let B be the fixed installment. The phrase 'premium refund' must be anchored to the policy: P may mean the single premium, a defined percentage of it, net consideration, or another specified basis. Let K = k indicate that exactly k scheduled life payments have been made before the insured death event. A simple residual is

Rk=max⁡(P−Bk,0).R_k=\max(P-Bk,0).

This expression assumes every installment counts in full toward the refund threshold, no additional refund interest, no premium taxes or refunds already paid, and no partial-payment corrections. Many products need different accounting. At Bk ≥ P, the death-refund amount is zero. The formula counts paid benefits, whereas contract language might instead count amounts 'payable' or 'credited'; that semantic choice changes the death-administration ledger.

Under an annual cash-refund convention, death occurs during year k+1 after k completed payments; the cash refund R_k is disbursed at the end of that year. The expected present value of the death benefit is therefore

Crefund=∑k=0∞vk+1(Sk−Sk+1)max⁡(P−Bk,0).C_{\mathrm{refund}}=\sum_{k=0}^{\infty}v^{k+1}(S_k-S_{k+1})\max(P-Bk,0).

The total level-payment pure-life-plus-cash-refund value is B a_life + C_refund. This is a model valuation formula, conditional on a stated death-settlement lag. Real policies can pay on receipt of due proof of death, producing a potentially different random settlement time; there is no basis for treating administrative delays as cost-free. A positive refund floor is not a cash surrender value exercisable by the annuitant while alive.

Installment settlement and the equivalence boundary

An installment-refund guarantee continues scheduled installments to a beneficiary after the annuitant's early death until the cumulative defined payments meet the premium-refund benchmark. Suppose P/B = N is a positive integer, with exactly N annual installments as the target. Every first-N payment is owed to the annuitant while alive or to a beneficiary after death. Consequently, on the specified annual payment convention, the installment-refund contract is cash-flow equivalent to a life annuity with N payments certain. Its present-value factor is a_(N-certain+life). This exact equivalence is conditional on equality of beneficiary priority, the definitions of credited payments and refundable premium, and the installment dates; it is a useful contractual test rather than a blanket assertion about products with similar marketing labels.

If P/B = n + f, where n is a nonnegative integer and 0 < f < 1, the settlement generally involves n full credited installments and a final fractional residual fB if death intervenes before recovery. In the synthetic end-of-year convention, payments 1,...,n are guaranteed at B, and at time n+1 a payment of B is made if the annuitant survives or fB is paid if the annuitant has died before that date and is otherwise entitled to the residual. Its expected factor is

ainstallment=∑k=1nvk+vn+1[Sn+1+f(1−Sn+1)]+∑k=n+2∞vkSk.a_{\mathrm{installment}}=\sum_{k=1}^{n}v^k+v^{n+1}[S_{n+1}+f(1-S_{n+1})]+\sum_{k=n+2}^{\infty}v^kS_k.

This formula assumes the fractional residual is settled on the same scheduled date rather than immediately after the particular death date, and assumes the life annuity remains payable in full if survival occurs at the n+1 payment date. Alternative policy wording requires a separate path calculation. If f = 0, the expression collapses to n-certain-plus-life. If mortality is zero through the recovery horizon, the guarantee has no extra contingent value, though the insured still receives the stated life payments.

Cash and installment timing: a controlled comparison

Take P = $50,000, B = $10,000, and death after exactly two annual installments have been received. The refund balance is $30,000. Under an illustrative contract with payment at the next year-end, cash refund pays $30,000 at time 3. An installment refund instead pays $10,000 at times 3, 4 and 5. With 3% annual discounting and present value evaluated at time 2, the cash settlement is $30,000/1.03 = $29,126.21, whereas the remaining installments are $10,000[(1.03)^(-1)+(1.03)^(-2)+(1.03)^(-3)] = $28,286.11. The cash arrangement has a greater present value by $840.10 in that particular death state. The two alternatives generate the same $30,000 undiscounted cumulative amount; the difference is time, liquidity and reinvestment opportunity.

A positive discount rate and identical death-triggered refund amounts imply an earlier cash payment is at least as valuable in discounted terms as later installments, provided beneficiary entitlement and credit risk are held constant. If the insurer prices a fixed initial premium endogenously, however, annual payout B changes across options. The ratio P/B and the duration of guaranteed recovery then also change. A clean market comparison must solve each contract's full pricing equation; it cannot hold the same benefit B, the same premium P, and the same zero-loading value simultaneously across all alternatives.

Under the synthetic mortality and interest basis used earlier, a life annuity with B = $10,000 and five-year installment refund has modeled present value $125,059.16. A cash-refund arrangement with P = $50,000 and B = $10,000 has modeled life-payment value $118,750.00 plus death-refund value $6,570.06, or $125,320.06. The cash refund is $260.89 more expensive under those specific benefit and death-settlement conventions. These two packages are compared at fixed B and fixed reference P, not at a self-consistent equal gross-premium quotation. The figures show a cash-flow ranking, not a price menu.

Pricing equations and the nonlinearity of a premium benchmark

Let G denote the gross single premium, E(B,P) the present value of acquisition and maintenance expenses, and M(B,P) the chosen cost of capital, profit and other provisions. A generic cash-refund pricing relation is

G=Balife+Crefund(B,P)+E(B,P)+M(B,P).G=B a_{\mathrm{life}}+C_{\mathrm{refund}}(B,P)+E(B,P)+M(B,P).

If the refundable reference is itself G, the refund term depends on the premium being solved for. If the premium is fixed and B is quoted, the refund term depends on B both through the residual and through the duration for which it remains positive. The pricing problem is piecewise rather than a simple multiplication by one constant annuity factor. When P/B passes an integer, a payment period enters or leaves the refund exposure. An algorithm should keep track of this threshold rather than approximate it with a permanently fixed guarantee term.

An actuarial tariff requires a mortality basis appropriate to annuity purchasers, discount assumptions appropriate to its pricing objective, timing of claim settlement, premium loads, investment and operational risk, and contract options. The IAA's actuarial standards directory identifies general actuarial practice and IFRS 17 actuarial work standards relevant to governance, although jurisdiction-specific applicability depends on professional adoption. Pricing, financial reporting, and solvency valuation need separate bases and documented reconciliations.

Early-Death Value Transfer

Beneficiaries, estates, and who owns the promise

A guarantee redistributes early-death entitlements. Under a straight life annuity, the cessation of future lifetime payments is part of the pooling arrangement. Under a period-certain guarantee, the residual series is owed to a designated contractual recipient after death within the guaranteed term. Under cash refund, a defined lump-sum death benefit becomes due if the paid benefits have not yet reached the reference amount; under installment refund, residual installments are owed. Product forms need to specify beneficiary designation and change, contingent beneficiaries, successor ownership if relevant, proof of death, whether commutation is permitted, and treatment of payments issued near the date of death. The Compact's beneficiary rules explicitly address beneficiary designation and effectiveness, reinforcing the contractual priority of the operative document.

An early-death transfer is the insurer's payment of a defined benefit according to a policy. It should not be equated with the annuitant's retained investment account or a full inflation-adjusted recovery of contributions. A refund priced as part of an annuity can have an actuarial cost even though its expected use is conditional on early death. Beneficiaries may value a lump sum more highly when liquidity is needed urgently; others may prefer installment discipline or a predictable income schedule. These preferences are not directly measured by the valuation identity.

Economic incidence and mortality credits

For a fixed premium and consistent assumptions, adding a refund or guaranteed-period obligation uses funds otherwise available for survival-contingent annuity income. The reduction in payment rate can be understood through the valuation decomposition: more expected value is assigned to death states within the guarantee horizon. The life annuity still insures survival beyond an unexpectedly long age, but it may do so at a lower regular payout. The effect is a transfer between risk states and beneficiaries, mediated by the insurer's pricing. It is not proof that every annuitant with dependents should purchase a refund option.

The welfare question belongs in a larger retirement portfolio. Davidoff, Brown, and Diamond (2005) demonstrate that market incompleteness and consumption needs can change optimal annuitization outcomes. The OECD payout-phase discussion stresses that essential living expenses, public pension benefits, liquidity and health circumstances interact. A person with other resources for bequests may value maximizing lifelong payments, whereas a household with dependent beneficiaries may value a death-state guarantee. The monograph identifies the mechanisms and does not rank household products for any particular consumer.

Evidence about selection and heterogeneous purchasers

Economic valuation and observed product demand diverge when buyers have private information about health, preferences, or family obligations. The seminal Finkelstein–Poterba annuity study analyzes British insurer policyholder data and reports systematic relationships between realized mortality and contract characteristics, including options affecting estate payments. It also reports no substantive mortality difference by annuity size in that analysis. These findings caution against assuming that one universal annuity mortality table prices every guarantee option accurately. They do not establish a causal effect of selecting a refund feature on mortality, nor do they directly estimate contemporary U.S. pricing differentials.

The SOA–LIMRA payout annuity study describes 2020–2024 individual payout annuity mortality observations from 23 parent-company groups and 26 contributing companies, representing over 80% of industry sales during the study period, with approximately 3.1 million contract-years of exposure and 143,190 deaths. The public summary supports the availability and broad scope of recent annuitant mortality experience. The study's paid data package and detailed segmentation were not obtained in this research; hence no option-specific mortality rate or credible empirical estimate of refund-buyer selection is asserted here. The study distinguishes evidence on actual portfolio mortality from the deliberately artificial survival schedule used in the numerical examples.

Tax attributes are separate from benefit design

An economic refund of nominal premium and a tax-law recovery of investment in the contract are different concepts. For the United States, IRS Publication 939 describes the General Rule for nonqualified annuity payments and, for certain qualifying arrangements, adjustments to investment in the contract for the value of a refund feature. Its definition of a refund feature involves lifetime-contingent expected return, a payment after annuitant death when a specified amount or number of payments has not been received, and a refund of the contract cost. The publication uses specific IRS actuarial tables and statutory categories. A household's stated premium guarantee cannot by itself determine its exclusion ratio.

IRS Publication 575 addresses taxable and tax-free portions of pension and annuity payments, different treatment of qualified and nonqualified arrangements, and special rules for guaranteed payments received by beneficiaries. It explains a particular beneficiary rule under which guaranteed distributions can initially be excluded until the cost recovery threshold is reached, subject to the rule's limits and exceptions. The detailed treatment turns on the tax status of the plan, annuity starting date, applicable calculation method, the annuitant's prior exclusions, and how the beneficiary receives payments. This analysis reports the 2025 federal publications in force as available in November 2026, rather than presenting a universal tax result or individualized tax advice.

Cross-jurisdictional and accounting boundaries

Legal standards arise from jurisdiction-specific combinations of insurance contracts, beneficiary law, annuity regulation, pension law and taxation. The U.S. Insurance Compact standards identify conditions for their product-filing scope, while the OECD DC recommendation sets policy-level considerations for retirement plan design. Neither may be substituted for another jurisdiction's contract form or taxation regime. Cross-border comparisons should start with the product's actual benefit definition, currency, applicable insurer authority, consumer disclosures, and legal enforceability.

A further distinction concerns accounting liability versus economic market price. The IFRS 17 measurement overview requires measurement of groups of in-scope insurance contracts using fulfillment cash flows and contractual service margin where applicable, with identification of distinct investment components and other obligations. Death-triggered refunds and guaranteed installments can alter projected cash flows and contract service assessment. Recognition and presentation require analysis of the actual insurance risk, contractual terms, and IFRS 17 classification; a generic refund label is insufficient. The insurer's ability to meet obligations remains a prudential question assessed under the applicable solvency framework.

Evidence Validation and Model Sensitivity

Mortality basis, data lineage, and selection

A modeled guarantee requires an explicit survival basis by sex or other appropriate classifications, issue age, calendar period, duration from purchase, annuity type and benefit election. Annuitants are generally selected differently from the population at large; contract features can introduce further heterogeneity. A rate fitted to life-only annuities may not transfer without adjustment to cash-refund purchasers. The SOA's 2026 experience study supplies a recent institutional starting point for U.S. payout-annuity mortality, while Finkelstein and Poterba (2004) motivate a test by product attribute. The present work does not download insurer microdata, fit a survival model, estimate selection parameters, or assert market prices. Such work would require disclosed exposure definitions, deaths by duration and option, and validation against held-out experience.

Data must distinguish annuitant deaths from other contract terminations and resolve partial periods. A death-notification delay can make a payment appear to continue after death; the analyst must reconcile overpayments, recoveries and contractual benefits. Observed payments to beneficiaries cannot be treated as annuitant survival. The clean empirical unit is the entitlement at each payment date, tagged by its legal trigger and actual recipient. This prevents administrative events from contaminating the survival curve and refund recovery ledger.

Interest-rate and payment-time sensitivity

For positive future cash flows, increasing the discount rate decreases their present value when everything else is held fixed. A guarantee that adds short-duration installments generally changes cash-flow duration relative to a pure life annuity. The direction of the difference between two payout-rate quotes also depends on their changing bases. Cash refund concentrates a conditional death payment sooner than installment refund; for equal nominal death amounts and a positive discount rate, earlier settlement has higher present value. That comparison presumes equal credit quality and ignores taxes and beneficiary preference effects.

The interest term structure matters because long-tail survival payments and short-horizon death refunds occupy different maturities. Replacing a nonflat curve with a single average rate may alter option relativities. The identity for a period-certain factor survives under term-specific discount factors D_k, provided D_k is applied consistently:

Δam=∑k=1mDk(1−Sk).\Delta a_m=\sum_{k=1}^{m}D_k(1-S_k).

The analysis can be extended to scenario-dependent discounting, but correlation between mortality, inflation and market returns may then make the expected product of stochastic discount factors and survival indicators relevant. A point-estimate mortality curve combined with a deterministic discount curve does not quantify that dependence.

Solvency, accounting and claims expenses

In the European Union, Solvency II Article 77 sets technical provisions as the best estimate plus risk margin and describes the best estimate as the probability-weighted present value of future cash flows using the relevant risk-free interest-rate term structure. A cash refund is a prospective death benefit and an installment refund creates surviving beneficiary payments; both belong in projected obligations where relevant. Expenses for beneficiary verification, payment processing, records maintenance and potential disputes can differ by product design. The overall technical provision follows the operative prudential basis, not the simplified zero-expense example.

IAIS ICP 14 on valuation requires transparent, reliable valuations for solvency and explicitly explains the need to reconcile general-purpose and regulatory reporting, including discount rates, methods and assumptions. IFRS 17 follows its financial-reporting measurement objectives. A single insurer may therefore produce different, internally coherent valuations of the same promise for pricing, accounting, and solvency. Contractual entitlements remain the common cash-flow foundation while measurement purpose governs assumptions, margin structure and presentation.

There is a time-sensitive regulatory issue: the 2026 EIOPA guideline update states that revised risk-margin guidelines become applicable on January 30, 2027, while prior guidelines apply until then. The present November 2026 analysis uses the currently described Article 77 principle without retroactively applying the 2027 guideline. The same effective-date discipline applies to the Insurance Compact's December 2026 amendment.

Sensitivity protocol and independent mathematical checks

The model is reproducible from the following input set: scheduled installment dates, B, contractually defined P, guarantee term m, survival sequence S_k, settlement rule, discount factors, expenses, and projection horizon. The minimum exact checks are (1) S_0 = 1 and 1 ≥ S_k ≥ S_(k+1) ≥ 0; (2) the life factor is the survival-weighted sum; (3) m = 0 recovers the life factor; (4) adding one guaranteed period increases value by D_(m+1)(1−S_(m+1)); (5) a five-payment installment refund with P/B = 5 recovers the five-certain factor; (6) a cash refund of P after zero prior payments satisfies its maximum contractual residual; (7) once cumulative credited benefits exceed P, the death-refund is zero; and (8) for a positive discount rate and otherwise identical early-death rights, earlier cash settlement dominates the corresponding later installments in discounted value.

The synthetic illustration uses S_k = 0.95^k and D_k = (1.03)^(-k). Its exact geometric life factor is 11.875. The five-certain factor is 12.5059162145. The cash-refund expected present value for P = $50,000 and B = $10,000, with death settlement at the next year-end, is $6,570.0564, so total cash-refund value is $125,320.0564. At the same B, the five-certain or five-installment-refund value is $125,059.1621. A death after two installments produces residual $30,000. These independently recalculable values give falsifiable tests of an implementation rather than an empirical estimate of any insurer's portfolio.

Sensitivity findings follow from cash flows: longer guaranteed terms weakly increase the actuarial cost at fixed B; a higher discount rate lowers the present value of positive payment obligations; and lower early-period survival generally increases the incremental period-certain guarantee component. The total life-plus-guarantee effect of mortality changes is not captured by the increment alone, and mortality-model updates may affect future payments beyond the term. Nonlevel inflation-linked annuities require indexed B_k, price-index definitions and a new residual-crediting rule; separate simulation and contract evidence are necessary before reporting quantitative results for those products.

Source verification and restrictions

Sources were reviewed at the level of their publicly accessible official text, publisher bibliographic records, or explicit study summaries. The SOA 2020–2024 detailed data package is restricted and was not reviewed; no claim here depends on its undisclosed option-level estimates. The historical economic literature is used for established theoretical propositions and documented observational findings, with paywalled portions treated as abstract-level evidence where applicable. The numerical model is original derivation, not a description of a named insurer's filed pricing manual. The present evidence ledger documents a source-to-claim mapping, access limitations, and verification dates. Particular contract wording, applicant mortality selection, institutional premiums, settlement practices, and actual tax classification must be separately established for any applied valuation.

Integrated Research Findings and Open Scientific Questions

A unified interpretation of guarantee mechanics

The central finding is that a guarantee is best understood by the mapping from survival and death states to amounts, recipients and due dates. Period-certain guarantees introduce a fixed schedule over a specified number of dates; cash-refund options introduce a death-triggered residual amount; and installment-refund options convert a premium recovery benchmark into a scheduled residual series. Under an integer recovery horizon and matched conditions, installment refund becomes pathwise equivalent to a period-certain term. Cash refund preserves the same defined nominal death-state recovery threshold while accelerating settlement. These results are algebraic consequences of the contracts and timing assumptions, and no novel named theory is needed to explain them.

The second finding is that premium comparison requires separating nominal benefit, expected present value and actual insurer pricing. Under fixed B and P, stronger early-death payment rights have weakly larger present value if all other bases remain equal. Under fixed G, the insurer may lower B, change the effective P/B horizon, or impose different fees. Contractual value cannot be inferred from a marketing phrase or from one benefit component. A refundable 'premium' may be a net contractual amount, and nominal cost recovery does not promise preservation of real value. Full comparison must disclose which variables are held constant.

The third finding concerns uncertainty and distribution. Early-death guarantees protect a beneficiary's claim but reduce the amount of purely survival-contingent pooling available for a given financing base. The appropriate retirement-income trade-off depends on public and private income, household consumption, longevity exposure, bequest motives, and liquidity. Institutional consumer guidance, OECD retirement-income analysis, and the annuitization welfare literature jointly support an assessment based on the household's complete economic setting. No universal recommendation follows.

The fourth finding is methodological: mortality selection, claim administration, and legal definitions are first-order inputs. A study using aggregate population mortality, an implicit death-payment delay, or an ambiguous premium benchmark may appear numerically precise while answering the wrong contract question. Current SOA payout-annuity experience provides the direction for stronger mortality evidence, but accessible summary data do not resolve selection by specific guarantee election. Research should collect option-level benefit terms, insured exposures, payment dates and death notifications before fitting product-choice or cost models.

Comparative research agenda

A fully empirical study should assemble versioned specimen contracts or filed forms across representative jurisdictions, with exact wording on guaranteed term, refundable base, claim maturity, commutation, credited payments and beneficiary priority. It should then match issued-policy cohorts by age, underwriting, purchase timing, premium and option choice, documenting each source's permissions and data provenance. Modeling should compare deterministic and stochastic mortality, yield curves, expenses, settlement lags and guarantee election behavior, without importing untested parameters from a different insurer or decade. Financial reporting and solvency comparisons must be anchored to each jurisdiction's reporting and capital measurement rules.

Three especially valuable open questions follow. First, how large is contract-feature selection after conditioning on observable health and purchase channels? Second, how much of the observed cash-versus-installment option price reflects discount timing rather than capital costs, behavioral selection, commissions, or operational practice? Third, how do legal rights to commutation, deferred income commencement, inflation linkage and tax treatment change beneficiary welfare across jurisdictions? Each question is empirically testable with appropriate data and does not require a new theoretical doctrine. A source-critical design would state counterfactuals, event-time definitions and sensitivity ranges before estimating effects.

Research contributions and limits

This monograph's original contribution is an integrated source-verified comparison of the three guarantees, a fully stated discrete-time valuation convention, a derivation of exact integer-recovery equivalence and conditional fractional-recovery expression, a cash-refund valuation formula with explicit death settlement timing, and a set of independently reproducible illustrative checks. Its comparative interpretation connects actuarial cash flows to consumer protection, beneficiaries, tax classification, IFRS reporting and prudential valuation. It claims no discovery of new actuarial laws, no causal inference from unavailable data, and no original rate table. The unresolved issues are questions for future empirical studies, not barriers to validating the carefully delimited claims and formulas completed here.

Conclusion

Period-certain and premium-refund guarantees are distinct legally enforceable designs for distributing the value of life-contingent retirement income across survival and premature-death outcomes. The period-certain rider secures a minimum schedule and preserves lifetime payments beyond it. Cash refund provides a defined early-death residual in a single settlement, while installment refund spreads recovery across scheduled dates. Under stated conventions an integer installment-recovery horizon coincides with a life-plus-certain guarantee, whereas accelerated cash settlement changes present value even when nominal recovered amounts coincide. Each promise requires precise identification of benefit triggers, credited payments, recipients, settlement dates, and contractual premium base.

The actuarial equations establish the direction of elementary cash-flow effects and show the conditions under which comparisons are valid. Economic value additionally depends on selected annuitant mortality, interest rates, expenses, capital, inflation, beneficiary liquidity and insurer credit. Legal entitlement, tax cost recovery, accounting measurement and solvency reserves address separate questions. The primary documents and research reviewed establish the institutional relevance of those distinctions, while the synthetic examples provide testable calculations with fully declared assumptions. A scientifically reliable assessment proceeds from original wording and a dated jurisdictional source through a replicable contingent-payment model to a clearly scoped interpretation of evidence. The dossier is complete as a foundational analytical study at the November 15, 2026 research cutoff; empirical option-level estimation remains an independently specified research task.

References