RESEARCH DOSSIER
Deferred Income Annuity Deferral and Survival Vesting
Abstract. A deferred income annuity (DIA) exchanges a premium, or a defined series of premiums, for contractually specified future income that ordinarily depends on the survival of an identified annuitant to a deferred payment date and, for life-contingent installments, to subsequent payment dates. The deferral interval changes the distribution of benefit receipt, the mortality credits available to survivors, the duration and sensitivity of insurer liabilities, and the economic significance of guarantees for deaths before commencement. This monograph analyzes deferral and survival vesting as a connected contract-design problem. It establishes the necessary distinctions among an executed right to a conditional future promise, the satisfaction of a survival condition, the accrual of a particular payment, and a separately promised beneficiary benefit. It develops discrete and continuous actuarial valuation identities, examines refund and survivor provisions, evaluates timing and option risks, compares selected regulatory regimes, and reviews contemporary empirical evidence on annuity pricing and mortality selection. A fully reproducible stylized calculation separates the effect of deferral, survival assumptions, discount rates, and a pre-commencement premium refund. The principal finding is that the deferred payment date is simultaneously a contractual trigger, a mortality-selection boundary, a household liquidity commitment, and an insurer valuation horizon. The term survival vesting is used with an explicit analytical definition; the legal classification of rights remains jurisdiction- and contract-specific. No new actuarial theory or empirical population estimate is claimed.
Keywords: deferred income annuity; longevity annuity; survival contingency; survival vesting; annuity commencement; deferral period; mortality credits; annuitant mortality; refund of premium; contingent beneficiary rights; life-contingent valuation; longevity risk; QLAC; retirement decumulation; actuarial present value; retirement income guarantees.
Introduction
Deferred income annuity contracts organize financial protection around a future date at which life-contingent income is scheduled to begin. An individual purchasing coverage at age 65 may choose commencement at age 75, 80, or 85. Payments may commence only upon survival to the specified event and may continue for life, subject to the precise payment calendar. A survivor who satisfies the contractual condition holds an enforceable benefit according to the contract. A person who dies before commencement may leave a refund or survivor entitlement if that separate promise was purchased. These distinctions determine the legal and actuarial substance of the arrangement.
The U.S. Securities and Exchange Commission's investor education materials characterize a longevity annuity as a fixed annuity promising future lifetime payments if the purchaser survives to the payment date, with a possible optional death benefit. This describes the pure contingent structure at the heart of this dossier. Broader deferred-annuity terminology also includes products with an accumulation account and a later, optionally exercised annuitization decision. The SEC's separate overview of deferred annuities discusses that accumulation-to-payout architecture. Their distinct cash-flow and entitlement mechanisms must be preserved during product comparisons.
The analytical problem has four parts. First, a deferral period is a specified interval between the contractual purchase or valuation date and scheduled income commencement. Second, the life insured, the payment recipient, the owner, and the beneficiary can be different parties. Third, survival conditions alter both the expected cost of benefits and the distribution of losses borne by purchasers who die before income begins. Fourth, contractual guarantee and refund clauses determine what, if anything, passes to a beneficiary despite failure of the principal survival condition.
The argument proceeds from primary contract and regulatory descriptions through established actuarial mathematics and international valuation frameworks to empirical evidence. It uses the OECD pension design recommendation, which recognizes the role of partial, deferred, and delayed lifetime income in retirement arrangements, as a policy comparator rather than as a universal contractual rule. Actuarial and accounting standards are interpreted within their stated scope. Source-specific restrictions are preserved: an empirical estimate from U.S. retail products is not generalized to every annuity market; a legal condition attached to a U.S. qualified longevity annuity contract is not attributed to all private deferred income annuities.
This is a scientific investigation of contract rights, mortality-contingent cash flows, and institutional risk, not a personalized product recommendation. Its calculations are stylized analytical demonstrations and are not insurer quotes.
Scientific Foundations, History, and Research Identity
The object of a deferred income annuity
Let the purchase date be time zero, let the life aged x be the covered annuitant, let d be the number of years until scheduled commencement, and let B be a stated periodic income amount. A pure deferred life annuity promises B on specified payment dates after time d while the annuitant satisfies the contractual survival rule. The insurer receives an upfront premium P or another contracted premium sequence. In the pure form, payment on death before commencement is zero. A refund or guaranteed benefit changes that form and must be modeled separately.
The economic reason for buying such an income stream is the transfer of uncertain late-life consumption duration to an insurer pooling many lives. A long-lived individual receives payments beyond the duration that their own earmarked assets might otherwise support. For a purchaser who dies before commencement, the premium finances the collective payments to survivors and insurer expenses, capital, and profit. The resulting mortality credits help support the level of promised lifetime income. This mechanism makes the survival condition integral to pricing.
The analytical label "survival vesting" designates satisfaction of the defined survival condition relevant to payment entitlement. It is not asserted to be a uniform statutory term, a transferable vested pension interest, or the same event as the signing of the insurance contract. The insurer's promise can be binding at issue even though one or more payments remain contingent. Contract formation, a protected interest in the policy, qualifying survival, and the maturity of an individual installment are distinct legal and economic events.
These distinctions also separate the DIA from a savings deposit, a guaranteed fixed-term income contract, a pension drawdown account, and a survivor-only death benefit. All can deliver cash flows in retirement. Only an identified life-contingent promise maps the due payment to the survival state in the manner examined here. The SEC's longevity-annuity description is directly informative on this point; by contrast, the National Association of Insurance Commissioners' general fixed-deferred-annuity buyer guidance addresses choices to annuitize an accumulation contract and does not establish every term of a pure DIA.
Historical and conceptual development
The scientific history of this product combines the older mathematics of life annuities with modern proposals to insure advanced-age consumption. Brown and Warshawsky (2001), in their study of longevity-insured distributions, examined how pension structures and access to annuity options affect retirement longevity insurance. Their analysis concerned the transition toward defined contribution retirement saving and the institutional availability of insured payouts.
Milevsky (2005), in Real Longevity Insurance with a Deductible, developed the advanced-life delayed annuity concept as an instrument concentrated on very late-life payments. This work is particularly relevant to the research identity of deferral: later commencement may allow retirement resources to be devoted specifically to the longevity tail. Its illustration of possible payments depends on the paper's assumptions and is not a contemporary product quote. The proposal also demonstrates that installment-financed deferral, inflation adjustment, and high-age commencement can be combined within a single design.
The theoretical case for annuitization must be interpreted in conjunction with liquidity constraints. Davidoff, Brown, and Diamond (2005), in Annuities and Individual Welfare, analyze welfare conditions favoring annuitization and explain why incompleteness of markets can constrain the optimal share. A deferred survival-contingent benefit is especially sensitive to the future timing of household spending, medical expenditures, unexpected borrowing needs, and bequest preferences. The lifetime benefit and the restriction of pre-commencement access are two features of one commitment.
Parties, dates, and documentary identity
A source-critical evaluation begins with the issued policy form and all endorsements. At minimum, the record should specify the issue date, premium receipt date, effective date, annuitant identity, owner, beneficiary, selected commencement date, first scheduled payment date, frequency, payment-in-advance or payment-in-arrears convention, death-reporting mechanism, and any conversion rights. The contract must also show whether the amount B is fixed at issue, depends on subsequent contributions, or may be revised on an elected date.
The words "first payment at age 85" can conceal several different administrative conventions. A benefit may be payable on the anniversary on which age 85 is reached, on the first day of the next month, at the end of the first monthly survival interval, or on another contractually specified date. When the payment test requires survival at the start of a month, it differs from a benefit contingent on survival to the month's end. These differences alter the earliest payable installment and the treatment of deaths close to commencement.
A contract also requires an explicit reference life when the owner and annuitant differ. Death of the owner need not end coverage if the policy is measured on another life and the ownership interests pass under relevant contractual and legal provisions. The scientific question is whose survival activates the insurer's income liability, not merely who signed the purchase application.
Research boundaries and source discipline
The core findings use contractual descriptions, regulatory text, actuarial identities, and published study summaries. The mortality evidence of one population cannot automatically substitute for the rates of selected annuity purchasers. The Society of Actuaries and LIMRA's 2015–2024 deferred annuity study covers a broad deferred-annuity experience set; this source alone does not isolate pure DIA purchasers with identical income-start options. It is useful for evidence governance and selection concerns but does not license a DIA-specific mortality calibration.
A dated contract edition, issuing jurisdiction, and regulatory effective date are therefore primary research variables, alongside mortality base, cohort adjustment, discount curve, and guarantee elections. This discipline prevents terminology from substituting for evidence.
Deferral Period Choices
Definition and temporal architecture
Deferral is measured from a declared starting event to a declared commencement event. If the policy is issued at age x, a contract with commencement age a has a nominal age-based deferral of d = a − x. Calendar dates are needed to translate ages into payable dates. A month-level representation is often appropriate for retail annuities: if commencement is M months after purchase, the first qualifying monthly benefit may fall at M, M + 1, or another specified index depending on payment timing.
For life aged x, the unconditional probability of survival to time d is written as
where T_x denotes remaining lifetime at purchase. Under a deterministic force of mortality μ(x+t),
This survival probability is a property of an assumed mortality model. It is not a promise that a fixed percentage of actual policyholders will survive. The distinction matters when comparing a small insured pool with the expected value generated from national population experience.
The start date sets both the household's income bridge and the insurer's contingent payout horizon. During deferral, the purchaser must finance consumption without the deferred benefit. At commencement, the future stream begins according to the actual contract and continues only while the relevant survival rule holds, except where a separate guarantee applies.
Timing, money's worth, and mortality credit
For the same premium and a pure survival-contingent benefit, moving commencement later removes earlier payment dates and introduces survival filtering before the remaining benefit dates. Subject to a fixed discount and mortality basis, the present-value factor per unit of annual benefit decreases as the start date moves later. A level annual B priced to the same premium consequently increases. That mathematical comparison holds the product structure and assumed valuation basis constant. Retail quotations can also reflect different expense loadings, product availability, reserve charges, and underwriting bases.
The price difference has two sources: additional time discounting and additional pre-payment survival selection. Economically, the purchaser trades the possibility of receiving any benefit, immediate liquidity, and the length of potential payment service for a higher contingent benefit after commencement. A payout ratio calculated only for people alive at the commencement date omits the probability-weighted loss of benefit for those who die earlier.
This selection issue appears in Poterba and Solomon's money's-worth comparison. Their paper, issued in 2021 and revised in October 2025, analyzes May 2024 U.S. retail annuity quotes. For the deferred product purchased at age 65 and beginning at age 75, the reported expected present discounted payout per premium dollar is about 0.73 using U.S. population mortality and close to 1 using mortality representative of annuity purchasers. These are explicitly mortality-basis-dependent estimates for specified products and dates, not a universal rate of insurer profit or a promise about an individual's realized return.
Household choice and income bridge
The period before commencement requires a financing plan. Public pensions, employment income, bank deposits, portfolio withdrawals, and other insured benefits may cover the gap. The feasibility of that bridge affects the value of different commencement ages. Two otherwise identical households may evaluate the same DIA differently because one has sufficient income through age 85 and the other depends on liquidity in the preceding decade.
A later age concentrates insurance on the severe tail of longevity risk but leaves more intermediate-age consumption unsecured. An earlier age provides income sooner and across more states of survival at the cost of a lower amount per surviving payment under a common actuarial pricing basis. The OECD's lifetime-income design recommendation expressly includes deferred or delayed lifetime payments combined with programmed withdrawals as possible flexibility within pension payout arrangements. It also recommends mortality assumptions that reflect future improvement and clear communications about projected real retirement income.
This supports a conditional policy finding: commencement choices should be evaluated against total retirement resources and essential-spending obligations, not solely against the headline monthly income amount. It does not mean every person should choose a deferred annuity.
Timing options and their valuation
Some contracts lock d at purchase. Others allow a later election or modification before commencement. An option to accelerate payment onset, postpone it, or change the payout frequency carries economic value if it changes benefits in response to new health or interest-rate information. The realized value depends on contractual conversion factors, election windows, underwriting restrictions, and any reduction of monthly income.
An insurer must distinguish a truly fixed commencement promise from a policyholder option whose exercise alters expected cash flows. A contractual conversion formula may partly neutralize the financial effect but may leave mortality-selection incentives when an annuitant learns information about health after purchase. The purchaser may experience a substantial reduction in benefit when changing dates; the SEC alerts readers to date-change consequences.
The empirical research task for date options requires observing original election rights, election exercise, accrued benefits, and pre-election health or demographic information. Data limited to actual commencement dates cannot identify the original option set or disentangle date selection from survivorship.
Deferred Survival Contingency
States of entitlement
A mathematically useful contract-state description distinguishes four events: purchase and effective coverage, survival through the deferral interval, qualification for the first payment, and qualification for later life-contingent payments. Legal benefits may arise at each stage, but the cash-flow entitlement depends on the specified trigger. A premium refund, if provided, is a separately contingent death benefit before commencement. A survivor's pension after the annuitant's death is likewise a separate promise defined by its own lives, percentages, and activation conditions.
Let I(T_x > t) indicate that the covered life survives to the contractual payment time t. For level annual payments B at times d+1 through d+N, payable in arrears on survival to each date, the undiscounted random benefits are
Under deterministic discount factors D(0,t), the expected present value of the pure deferred income is
For lifelong payments, the sum is over all remaining relevant dates, with a limiting treatment of the mortality table at advanced ages. The model must document whether N is a contractual terminal age or a computational truncation and quantify any residual tail.
The formula emphasizes that entitlement to each installment is tested at its due date. Survival to commencement alone may be insufficient for payment after death, and a contract may make a beginning-of-period installment payable even when the annuitant dies during the period. Different payment timing alters the indicator and thus the expected present value.
Conditional versus unconditional valuation
At time zero the price contains both the chance of reaching commencement and the value of payments conditional on that event. For a fixed benefit design without pre-commencement payment,
provided the remaining conditional value is measured with the same valuation framework and date-d payment convention. The identity makes explicit why a benefit quoted only for survivors is not the average benefit value of all purchasers.
The surviving cohort at d will generally differ from the purchasing cohort at issue. Mortality heterogeneity can change the conditional population composition during deferral. Even if two purchasers have equal age and premium, their expected value may diverge with private health information. This is a central problem in applying published money's-worth estimates.
For a joint-life contract, a first-death or last-survivor rule requires a joint survival distribution. A single-life formula cannot be substituted without modifying the underlying state space. A beneficiary's receipt of payments under a guaranteed minimum period requires its own guarantee-state valuation rather than the continued survival of the deceased annuitant.
Mortality, interest, and expense bases
Mortality enters at least twice in a practical model: selection into the purchaser population and projected survival after issue. A base annuitant table may require cohort and calendar-year mortality improvement. Interest enters through a term structure or discount convention, potentially varying by accounting or prudential purpose. Expenses can include acquisition, policy maintenance, payment administration, beneficiary-claim handling, and guarantee administration. Lapse or surrender assumptions apply only if the issued terms actually permit such behavior.
There is no single universally correct discount rate for every purpose. Pricing, an economic comparison for a household, IFRS financial reporting, and a Solvency II best estimate serve different measurement objectives. Under IFRS 17, the insurance-contract measurement principles use current estimates of future cash flows, financial time value, risk adjustment, and contractual service margin as applicable. The simplified benefit formula above is an actuarial component and does not purport to replace the complete accounting measurement.
A correct model must prevent double counting. An expense already deducted from projected investment returns should not be added again as a distinct cost. A contractual refund should not be valued both as a contingent death payment and as an unconditional investment component without applying the actual contract terms.
Pre-commencement death and refunds
In the pure contract, death before commencement leaves no income benefit to the annuitant's estate. The SEC's description of beneficiary conditions makes clear that a beneficiary may receive nothing in the absence of an elected optional death benefit. Product wording, applicable law, and coverage form govern the actual outcome.
A return-of-premium-on-death rider changes the cash-flow vector. Let R(t) denote the death benefit due when death occurs at time t before d, and let f_x(t) be the modeled density of remaining lifetime. Ignoring claim settlement lag for the moment,
A constant refund R=P gives an expected present value that must be funded alongside the pure income stream. If a refund is payable at the end of the month following proof of death, its discount and administration timing differ. If the contract refunds premiums net of withdrawals, taxes, or riders, the modeled R(t) must reflect that definition.
At a fixed premium, adding a positive refund benefit ordinarily reduces the available pure income benefit, unless subsidized from another source or offset by other changed assumptions. The converse follows directly from cash-flow accounting: retaining a no-refund provision concentrates value on survivorship states and permits a larger survivor-contingent payment for the same modeled premium. Market prices need not equal this simple net-premium identity, but they are constrained by the same underlying obligations.
Failure modes in interpreting death protection
The terms "death benefit," "refund," "guarantee period," and "joint-life continuation" describe different contingent obligations. A premium refund that applies only to death before commencement does not automatically protect beneficiaries after commencement. A period-certain guarantee may protect installments for a set period after income starts, subject to the precise counting of elapsed installments. Joint-life continuation depends on another named life and may be paid at a percentage of the original benefit. The British public guidance on annuity death protections illustrates the practical distinction between the cessation of lifetime-only income and protection elected for beneficiaries.
The scientific record should state the event activating each benefit, its beneficiary, due date, calculation base, funding, and interaction with other riders. A schedule containing all claims payments is not enough without state-dependent activation rules, because differently contingent benefits may generate similar aggregate cash flows under a limited observation window.
Vesting and Commencement Design
Contract formation and payment conditions
The term vesting has multiple legal uses. Employment-pension vesting often concerns a participant's nonforfeitable entitlement in a pension program; insurance contract rights may instead concern an enforceable promise whose payment is conditional on future survival. In this monograph, survival vesting refers specifically to the factual satisfaction of the survival condition stipulated for the relevant annuity payment. This is a defined research usage, not a proposed universal legal doctrine.
At issue, the purchaser can acquire a legally enforceable claim to performance if future contract conditions are fulfilled. The future contingency does not make the existing insurance contract fictitious. Conversely, an enforceable conditional promise does not imply an immediate right to receive every scheduled payment. The separate concepts should appear consistently in contract drafting, consumer disclosure, valuation models, claims administration, and beneficiary communications.
A payment-calendar example demonstrates the problem. Assume a contract states that payments begin monthly on the first of the month following an annuitant's eighty-fifth birthday, in advance, provided the annuitant is alive on that first day. Survival at midnight on the first day makes the installment due subject to the contract. Death later that month may terminate future installments without reversing an already-due payment. A contract paying at month-end for survival to that date has a different outcome. This example is a logical comparison of hypothetical terms, not a claim about an actual insurer's issued form.
The United States: qualified longevity annuity contracts
A qualified longevity annuity contract is a particular tax-regulated subset of annuity design purchased through eligible U.S. retirement arrangements. The IRS Form 1098-Q instructions explain that a qualifying contract must meet conditions governing premiums, payment commencement, post-commencement distributions, and other provisions. The specified annuity starting date must be no later than the first day of the month following the employee's eighty-fifth birthday. Qualified status and required-minimum-distribution treatment depend on satisfying these criteria.
The 2024 final regulations implemented changes under the SECURE 2.0 Act: the former requirement limiting QLAC premiums to 25% of the account balance was removed, and the underlying statutory dollar cap was increased to $200,000 before specified inflation adjustments. The new final regulations apply to distribution calendar years beginning January 1, 2025, while their effective date is September 17, 2024. An author must distinguish the statutory base figure from any later inflation-adjusted applicable cap and must not treat a specific-year number as permanent.
The tax-regulatory ceiling is a qualification condition, not an actuarially optimal commencement age. A purchaser may choose an earlier permitted commencement. The absence of a cash surrender right after the relevant required beginning date and the rules for permitted death benefits affect product liquidity and beneficiary design. A qualifying contract may also allow the restricted 90-day rescission provision recognized in the 2024 changes. The exact boundaries should be determined from the applicable regulation and contract edition.
A conventional nonqualified DIA and a QLAC may share the same survival-contingent mathematics while differing in funding source, taxation, permissible contract options, reporting, and retirement-account integration. Therefore a statement true of QLACs cannot serve as a generic definition of all deferred income annuities.
United Kingdom: insured pension income and legacy terminology
British retirement guidance uses "deferred annuity contract" in contexts that are historically connected to pension buyouts as well as in the wider market for future pension income. The public pension buyout explanation describes Section 32 policies as insured arrangements preserving specified pension rights following a scheme transfer. The label in that context does not by itself establish a retail single-premium survival-only DIA with the same death and refund conditions as a U.S. longevity annuity.
The general guaranteed retirement income guide explains how annuity option selection can affect household income, beneficiary protection, and payments. It also distinguishes lifetime arrangements from other benefit forms. A sound comparison must therefore begin with the actual product's policy conditions and pension-law context, rather than equating every British use of "deferred annuity" with the U.S. QLAC definition.
This difference is significant for comparative research because ordinary language can obscure differences in the original pension rights transferred, the underwriting population, and the permitted age and commencement elections. A classification based only on the marketing phrase would generate non-comparable observations.
International pension design, accounting, and supervision
The OECD (2022) recommendation for DC plans promotes longevity-risk protection while allowing flexible combinations of deferred lifetime income and programmed withdrawals. The recommendation is an international policy instrument addressed to pension-plan design; it does not amend domestic insurance contract law.
International accounting treatment adds a second layer of classification. The IFRS Foundation's IFRS 17 overview describes insurance contract groups measured with risk-adjusted future cash flows and contractual service margin where applicable. Recognition and measurement depend on the applicable insurance-accounting boundary and contractual features, not only on the onset of the first annuity installment. Deferral may affect estimated benefit probabilities, risk duration, and profit recognition patterns without eliminating the existence of insurer obligations before commencement.
For European Union solvency purposes, Article 77 of Solvency II requires technical provisions to include a probability-weighted best estimate and a risk margin under the relevant rules. The EIOPA technical-provisions rulebook specifies expected future cash flows using the relevant risk-free term structure and current credible assumptions. A deferred income annuity can therefore create significant present liabilities before cash payments begin. The value of the obligation does not commence only when the first monthly benefit is paid.
EIOPA published revised valuation guidelines in 2026, with a stated applicability date of January 30, 2027. They address, among other matters, an adjustment to risk-margin methodology. This dossier distinguishes the 2026 publication of those guidelines from their later applicability date. They must not be described as already operational throughout 2026.
In Canada, the Office of the Superintendent of Financial Institutions' LICAT 2025 guideline specifically identifies longevity risk on annuities and prescribes risk-based capital calculations. Its longevity level stress differs by business registration and region, while trend risk reflects an adverse change in mortality-improvement assumptions. This is a jurisdiction-specific capital framework and should not be presented as a single global stress factor.
Vesting, data lineage, and operations
The operational chronology must be reconstructible from the underlying system. A reliable record contains the annuitant's verified birth date, actual issue date, contract conversion history, beneficiary elections, pre-commencement death notices, proof of survival if required, scheduled commencement event, and actual settlement. Data dictionaries should distinguish the scheduled first payment from the processed payment date and from the date on which the survival test was performed.
A delayed administrative payment does not automatically indicate that a contractual right arose late. Conversely, a scheduled date alone cannot prove the annuitant satisfied a survival contingency. Errors in recording these dates can distort observed mortality experience, reported reserves, and consumer entitlements. The International Actuarial Association's ISAP register identifies professional guidance for data, assumptions, documentation, and actuarial work. Local actuarial standards and mandatory legal rules must be checked separately.
Consumer-facing explanations should specify what the insurer promises before commencement, what happens on death during deferral, what qualifies a payment to begin, how later deaths affect installments, and which protections require additional premium or reduced income. Those elements directly control whether a household understands the practical consequences of selecting d.
Evidence Validation and Model Sensitivity
Distinguishing the available empirical populations
A critical empirical limitation is the difference between mortality of the total population, mortality of people who buy annuities, and mortality of people who both buy an annuity and survive to the deferred payout date. These groups differ by purchase choices and survival selection. Pricing studies that value identical payments using two mortality bases isolate one important consequence of that difference but do not identify all causes of observed pricing spreads.
Poterba and Solomon's U.S. quote-based analysis is directly relevant because it evaluates immediate and deferred individual products and reports results for differing mortality assumptions. Its revised 2025 research description relates to a May 2024 quotation sample. The reported 0.73 money's-worth figure for one deferred case is a specific estimate conditioned on its mortality and discount methodology, and should be assessed with the rest of its paper before being used in policy comparisons beyond the reported sample.
A different evidence object is the SOA/LIMRA deferred-annuity experience study, published September 17, 2026. The announced study covers 26 contributing companies, 2015–2024 experience, approximately 80 million contract-years of exposure, and over 1.4 million reported deaths. It reports broad deferred-annuity mortality experience rather than guaranteeing a separate, fully observed sample of pure DIA payment-commencement choices. The publicly announced high-level report was still described as forthcoming at the time of the source review; no unpublished mortality rate or DIA-specific effect is inferred here.
The U.S. Social Security Administration's 2023 period life table, published with the 2026 Trustees Report, is a transparent public benchmark. It describes age-specific mortality under period conditions, not forecast cohort mortality for future purchase cohorts. It should be used as an illustrative population comparator and not silently relabeled as an appropriate pricing table for an annuitant-selected portfolio. The consequences of mortality improvement during a long deferral interval may be substantial.
Gasimova, Haberman, and Millossovich (2025), in their deferred-annuity solvency study, examine deterministic and stochastic combinations of mortality and interest-rate assumptions for a deferred-annuity portfolio. The publisher identifies the paper as first published online in June 2024 and in a 2025 journal volume. Its abstract supports the need to consider the combined risks; numerical results not checked from the article are not reproduced.
Reproducible baseline calculation
The following constructed example tests the mathematics independently of quoted market prices. Assume: purchase age x = 65; a single premium P = $100,000; annual benefits in arrears; exact annual commencement delay d; 40 potential benefit dates following d; continuously compounded annual discount r = 3%; constant force of mortality μ = 2% at all modeled ages; no expenses, inflation adjustments, taxes, market spread, capital charges, or profit; and zero benefits on pre-commencement death. A constant hazard through advanced ages is deliberately a toy assumption and is not a realistic fitted mortality table.
Under the stated timing, the actuarial present-value factor for $1 paid at each qualifying date is
The net level annual income for the single premium is
For r + μ = a > 0, a computationally stable geometric-series expression is
The 40-date cap is an explicit computational endpoint, so for d = 20 it extends from age 86 to age 125. Such extreme ages underscore why this is a mechanical teaching example rather than a demographic forecast. The exact calculations yield:
| Deferral years | Survival to d under μ = 2% | Present-value factor A | Annual pure-life benefit from $100,000 |
|---|---|---|---|
| 0 | 1.00000 | 16.86456 | $5,929.59 |
| 10 | 0.81873 | 10.22888 | $9,776.25 |
| 15 | 0.74082 | 7.96626 | $12,552.95 |
| 20 | 0.67032 | 6.20413 | $16,118.30 |
The larger later-start benefits are conditional on reaching the payment dates. The table does not show annualized investment returns to all purchasers, and its survival column refers only to reaching d rather than to all later payments. These are distinct probabilities.
A precise independent reproduction can use the following pseudocode, with standard exponential arithmetic:
function factor(d, r, mu, N=40):
total = 0
for k from 1 through N:
total += exp(-(r+mu)*(d+k))
return total
income = 100000 / factor(15, 0.03, 0.02, 40)
This arithmetic is exactly equivalent to summing the discounted survival-weighted benefit indicators under a constant hazard. The mathematical identity does not require a proprietary model.
Mortality and interest-rate sensitivity
With d = 15, N = 40, and the same $100,000 premium, changing one assumption at a time gives:
| Scenario | Annual discount r | Constant hazard μ | Present-value factor | Annual pure-life benefit |
|---|---|---|---|---|
| Baseline | 3.0% | 2.0% | 7.96626 | $12,552.95 |
| Lower mortality | 3.0% | 1.5% | 9.23340 | $10,830.25 |
| Higher mortality | 3.0% | 2.5% | 6.89199 | $14,509.60 |
| Higher discount | 4.0% | 2.0% | 5.97845 | $16,726.75 |
| Lower discount | 2.0% | 2.0% | 10.73267 | $9,317.35 |
Each displayed annual benefit solves the net-premium identity within its own assumption set. These rows compare alternative issue-date pricing bases; they are not actual changes to an already guaranteed contract's benefit. Lower assumed mortality increases the present cost of every unit of promised survivor income and therefore reduces B for a fixed premium. A higher discount rate reduces present benefit cost and increases B in this deliberately simplified issue-date model. Real insurer pricing, asset yields, liability valuation, guaranteed rates, and capital requirements can respond in more complex ways.
For a fixed promised B, the liability direction is reversed: lower mortality or lower discount rates increases the actuarial present value of the promised obligation. The most important source of confusion in sensitivity analysis is failure to state whether B or P is being held constant.
Refund pricing sensitivity
Now add a death-before-commencement guarantee that pays exactly P at the moment of death during 0 ≤ t < d, and no refund after d. Under the same constant hazard and discount assumptions,
For P = $100,000, d = 15, r = 3%, and μ = 2%, the refund's model present value is $21,105.34. If the total net premium remains $100,000 and benefits are otherwise mutually exclusive, the amount available to finance life-contingent income under these idealized assumptions is $78,894.66. Accordingly,
This result is lower than the pure-life benefit of $12,552.95. It quantifies the economic cost of guaranteeing a payment in a state where the principal income promise would otherwise have paid nothing. The comparison assumes refund payment at death, a constant $100,000 benefit, no payment lag, no premium refund after commencement, and identical discount and survival assumptions; any real product's exact refund clause can change the outcome.
Sensitivity sources beyond the toy model
Realistic evaluation requires at least five additional adjustments. First, use age-, sex-, calendar-year-, and appropriately underwritten mortality assumptions, with explicit treatment of improving mortality and health selection. Second, represent cash flow monthly or at the contract's actual frequency and validate date conventions, including leap dates and first-payment offsets. Third, substitute a relevant yield curve and funding assumptions for one flat r. Fourth, attach expenses, premium taxes, guarantee fees, and the contract's actual risk adjustment. Fifth, test behavior only for contractual choices that exist: date changes, purchase cancellation, additional contributions, permitted commutation, and survivor-election changes.
Portfolio solvency requires sensitivity to simultaneous movements. The deferred-annuity solvency analysis provides an academic motivation for considering stochastic mortality and interest-rate risks together. The Canadian LICAT longevity requirements illustrate how a supervisor may separately calibrate level and trend longevity stresses. A deterministic arithmetic example demonstrates directions of exposure; it does not validate a firm's capital adequacy.
Source verification and inference limits
For the dossier's central factual claims, official regulatory or institutional texts were used to describe their own provisions. Publisher research pages were used to confirm study identity, dates, authors, and the conclusions expressed in accessible abstracts or public summaries. No full-text claim is asserted for a gated study on the strength of its title alone.
The source set covers the principal contract mechanics and selected U.S., UK, European, Canadian, and international contexts. It does not claim a complete census of all national annuity legislation or all issuer forms. The mathematical examples are transparent assumptions, not observations. The research discipline is to avoid treating the magnitude of a modeled premium refund, the share of annuity buyers surviving to commencement, or a regulator's stress factor as globally measured constants.
Integrated Research Findings and Open Scientific Questions
What the evidence establishes
The combined record supports six connected findings.
First, time deferral is simultaneously a consumption-timing decision and a survival-risk allocation. A later commencement age changes who receives money and when; it is a change in states of payment and not merely a delay in administrative processing.
Second, a valid insurance contract can exist before any payment is due. Surviving to a conditionally specified commencement date changes entitlement to the corresponding benefits, while the insurer's promise and prospective liability can exist earlier. This is the reason to separate survival vesting, contractual enforceability, and payment accrual in both legal analysis and actuarial models.
Third, the no-refund and refund designs distribute value differently across death states. The refund's economic cost can be expressed as an additional discounted death-contingent payment, making a reduced survivor-income benefit a direct consequence under a fixed net premium. Beneficiary rights should always be specified at the level of their trigger and payment calculation.
Fourth, the mortality basis is decisive for evaluating expected payout value. The empirical comparison by Poterba and Solomon provides a direct illustration of how population and annuitant mortality lead to different money's-worth results for the same quoted contract class.
Fifth, institutional labels do not create global legal equivalence. U.S. QLAC requirements, British deferred pension buyout vocabulary, European solvency valuation, and Canadian life-insurer capital calculations address different legal or supervisory objects. The IRS qualification rules, UK pension buyout guidance, and Solvency II valuation rule must each retain their jurisdictional scope.
Sixth, a scientifically defensible payout comparison requires a complete scenario definition: covered life, timing convention, mortality model, discount basis, benefit guarantee, source population, expenses, and product rights. A larger deferred payout is meaningful only in conjunction with its contingency structure.
Implications for research and disclosure
Product illustrations should place the deferral interval and pre-commencement death outcome next to the quoted income amount. They should also distinguish contractually guaranteed amounts from projections, disclose the first payable date precisely, and identify material date-change options. A projection for households should distinguish the probability of receiving at least one payment from the conditional size of payments once started.
Actuarial work should validate both issue-to-commencement survival and post-commencement payout duration, and should reconcile reserve cash flows with contractual benefit states. Mortality data lineage should include the difference between accumulation annuities and contracts whose payout amount is fixed on purchase. Data derived from broad deferred-annuity studies should be used with a documented product-composition adjustment or an explicit limitation.
International comparisons should identify tax qualification, pension-source restrictions, insurer solvency rules, and legal priority for beneficiaries. A universal use of the word vesting can otherwise merge employee pension accrual rights with the satisfaction of an annuity survival condition.
Open empirical questions
Several research problems remain open without preventing a complete synthesis of presently checkable evidence. One concerns the causal importance of private health information in commencement-age selection. Another concerns the frequency and distributional cost of pre-commencement death-refund elections. A third concerns administrative disputes around proof of survival and boundary-date deaths. A fourth concerns how calendar-year mortality improvement alters the valuation of contracts bought decades before payment. A fifth concerns how changes in discount curves, longevity forecasts, and guarantee take-up co-move during stress. A sixth concerns the relative treatment of flexible date changes under real issuer contracts rather than under hypothetical option schedules.
These are potential subjects for subsequent empirical research, not claims that additional proprietary theories have been established. Their resolution calls for de-identified issuer data, policy-form archives, careful contract-version matching, and repeatable actuarial tests.
Conclusion
Deferred income annuity deferral and survival vesting form a coherent research object centered on conditional future payment rights. The purchaser commits resources at an earlier date; the insurer becomes responsible for a specified contingent stream; survival through the deferral and subsequent payment periods determines which installments mature; and any pre-commencement death guarantee adds a separately defined obligation. The timing convention links all of these events.
Established actuarial present-value mathematics demonstrates why longer deferral ordinarily raises the net annual income available to qualifying survivors at a fixed premium and common valuation assumptions. The same mathematics explains the cost of refund provisions, the sensitivity to mortality and discount assumptions, and the difference between a guaranteed contractual benefit and its expected present value at issue. The illustrative model is exactly reproducible and its limits are explicit.
The international evidence establishes a further principle of scientific interpretation: the contractual survival state, financial-reporting obligation, pension qualification rule, and prudential capital measurement must each be examined under the governing source and jurisdiction. Current empirical studies improve understanding of mortality selection and annuity values, but the transfer of results between purchaser populations and product forms requires demonstrated comparability. These conclusions supply a tested analytical account of the deferred income annuity's design without asserting a newly created doctrine or an unverified claim of scientific novelty.
References
Brown, J. R., & Warshawsky, M. J. (2001). Longevity-Insured Retirement Distributions from Pension Plans: Market and Regulatory Issues. NBER Working Paper No. 8064. National Bureau of Economic Research. https://www.nber.org/papers/w8064
Davidoff, T., Brown, J. R., & Diamond, P. A. (2005). Annuities and individual welfare. American Economic Review, 95(5), 1573–1590. https://www.aeaweb.org/articles?id=10.1257%2F000282805775014281
European Insurance and Occupational Pensions Authority. (n.d.). Directive 2009/138/EC, Article 77: Calculation of technical provisions. Solvency II Single Rulebook. https://www.eiopa.europa.eu/rulebook/solvency-ii-single-rulebook/article-2160_en
European Insurance and Occupational Pensions Authority. (2026, July 15). Final report on revised guidelines on valuation of technical provisions—Solvency II Review. Guidelines stated to apply from January 30, 2027. https://www.eiopa.europa.eu/publications/final-report-revised-guidelines-valuation-technical-provisions-solvency-ii-review_en
Gasimova, K., Haberman, S., & Millossovich, P. (2025). Solvency analysis of deferred annuities. Decisions in Economics and Finance, 48, 853–871. First published online June 5, 2024. https://link.springer.com/article/10.1007/s10203-024-00452-2
IFRS Foundation. (2017, as amended and effective from 2023). IFRS 17 Insurance Contracts. Issued Standards. https://www.ifrs.org/issued-standards/list-of-standards/ifrs-17-insurance-contracts/
Internal Revenue Service and U.S. Department of the Treasury. (2024). Required Minimum Distributions: Final Regulations, T.D. 10001. Internal Revenue Bulletin 2024-33. https://www.irs.gov/irb/2024-33_IRB
Internal Revenue Service. (2025, April). Instructions for Form 1098-Q: Qualifying Longevity Annuity Contract Information. https://www.irs.gov/instructions/i1098q
International Actuarial Association. (2018 and subsequent updates). International Standards of Actuarial Practice: Register, including ISAP 1, General Actuarial Practice. https://actuaries.org/publications/international-standards-of-actuarial-practices/
Milevsky, M. A. (2005). Real longevity insurance with a deductible: Introduction to advanced-life delayed annuities (ALDA). North American Actuarial Journal, 9(4), 109–122. https://doi.org/10.1080/10920277.2005.10596229
MoneyHelper. (n.d.). Buy an annuity for a guaranteed pension income. UK Money and Pensions Service. https://www.moneyhelper.org.uk/en/pensions-and-retirement/taking-your-pension/guaranteed-retirement-income-annuities-explained
MoneyHelper. (n.d.). Pension buyout policies: An insurer pays your pension instead. UK Money and Pensions Service. https://www.moneyhelper.org.uk/en/pensions-and-retirement/pensions-basics/individual-buyout-policies
MoneyHelper. (n.d.). What happens to my pension when I die? UK Money and Pensions Service. https://www.moneyhelper.org.uk/en/pensions-and-retirement/pension-problems/pensions-after-death.html
Office of the Superintendent of Financial Institutions Canada. (2024, November 21; effective 2025). Life Insurance Capital Adequacy Test—Guideline (2025), including Chapter 6, longevity risk. https://www.osfi-bsif.gc.ca/en/guidance/guidance-library/life-insurance-capital-adequacy-test-guideline-2025
Organisation for Economic Co-operation and Development. (2022). Recommendation of the Council for the Good Design of Defined Contribution Pension Plans (OECD/LEGAL/0467). https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0467
Poterba, J. M., & Solomon, A. (2021, revised October 2025). New Evidence on the Money's Worth of Immediate and Deferred Individual Annuities. NBER Working Paper No. 28557. https://www.nber.org/papers/w28557
Society of Actuaries Research Institute and LIMRA. (2026, September 17). 2015–2024 Deferred Annuity Mortality Experience Study. Study landing page and public summary. https://www.soa.org/resources/experience-studies/2026/2015-24-deferred-annuity-mort/
Social Security Administration, Office of the Chief Actuary. (2026). Period Life Table, 2023, as Used in the 2026 Trustees Report. https://www.ssa.gov/oact/STATS/table4c6.html
U.S. Securities and Exchange Commission, Investor.gov. (n.d.). Annuities. Investor Bulletin and Educational Resources. https://www.investor.gov/introduction-investing/investing-basics/investment-products/annuities
U.S. Securities and Exchange Commission, Investor.gov. (n.d.). Longevity Annuity. Investor.gov Glossary. https://www.investor.gov/introduction-investing/investing-basics/glossary/longevity-annuity