RESEARCH DOSSIER
Insurance Service Coverage Units and Profit Allocation
Abstract. Insurance contract accounting transforms a contract's anticipated economic surplus into a sequence of reported service results. Under IFRS 17, the contractual service margin (CSM) represents unearned profit attached to a group of insurance contracts; the amount recognized in a reporting period follows the quantity of insurance contract services transferred during that period relative to current and expected future coverage units. This monograph investigates the legal and economic character of the service, the standard's measurement requirements, actuarial estimation of duration and benefit quantity, the interaction of insurance and investment services, and the consequences for profit allocation, comparability, disclosure, and auditability. The research distinguishes the service measure from expected claims, risk-adjustment release, premium cash receipts, and the broader CSM roll-forward. It examines the authoritative text of IFRS 17, its 2020 amendment process, the IFRS Interpretations Committee's 2022 annuity decision, actuarial professional guidance and peer-reviewed analysis. Six transparent computational experiments establish how flat, declining, revised, discounted and mixed-service patterns alter the timing of recognized CSM while preserving accounting identities under the stated assumptions. The central conclusion is that coverage units operationalize a service-based view of earned insurance profit, while material professional judgment remains in identifying the service and selecting a defensible measure of its quantity. Consistent definitions, documented assumptions, reproducible service projections, and disclosures linking the selected measure to the contract promise are therefore essential to meaningful cross-period and cross-entity interpretation.
Keywords: IFRS 17; insurance contract services; coverage units; contractual service margin; CSM release; profit recognition; general measurement model; variable fee approach; investment-return service; investment-related service; annuities; actuarial assumptions; financial reporting; service quantity; disclosure; financial statement comparability.
Introduction
A long-term insurance contract joins payment, uncertainty, protection, and time. The insurer can receive consideration before it fulfills years of service; the policyholder can receive protection in a year in which no claim occurs; and the contract may simultaneously provide financial investment services. This structure creates the accounting problem of identifying when a profitable contract group has earned the portion of its unearned margin that should appear in the period's financial performance.
The starting point is the IFRS 17 Insurance Contracts standard, issued by the International Accounting Standards Board (IASB) in 2017 and effective for annual reporting periods beginning on or after January 1, 2023. IFRS 17 establishes current fulfillment measurement and recognition of unearned profit over the period in which insurance contract services are supplied. At its most important level, the framework treats insurance as an obligation to provide protection and other qualifying services across time. The time at which premium is collected and the time at which an insured event materializes answer different questions from the time at which the insurer performs its service.
Coverage units are IFRS 17's quantitative expression of service provided by a group of insurance contracts. They allocate the remaining CSM between service already transferred during the reporting period and service expected in future periods. The operative B119 requirements in AASB 17, an official national adoption of the IFRS framework, provide directly accessible authoritative wording. The distinction between primary accounting text and complementary implementation literature matters throughout: professional guidance can illustrate acceptable methods and unresolved judgments, but does not replace the standard.
The investigation asks three questions. What counts as an insurance contract service, and what evidence demonstrates that the service exists? How can the quantity of that service and its expected duration be measured without mistaking the cost of claims for the benefit of protection? How does a defensible coverage-unit profile shape reported profit, and how can an external reader evaluate the resulting pattern? To answer them, this dossier conducts textual analysis of accounting requirements, compares official interpretations and peer-reviewed contributions, develops formula-based examples, and audits the assumptions that connect contractual promises with reported amounts. No claims of an original new named theory, compulsory universal calculation method, or empirical estimate of industry-wide profit effects are made.
Scope and disciplinary position
The principal scholarly home of this dossier is Insurance Accounting and Financial Reporting. The contractual structure of protection connects it to Life Insurance, Annuities and Retirement Insurance, Health Insurance, and Actuarial Science. Coverage-unit construction is examined at the intersection of accounting recognition, actuarial survival and persistency modeling, product design, economic service valuation, and financial-reporting governance.
Two neighboring topics have independent editorial boundaries. The full accounting mechanics of initial CSM measurement, interest accretion, assumption changes, currency conversion, and CSM reconciliation belong to Contractual Service Margin Accounting and Profit Emergence; they enter this analysis only to locate the amount available for allocation. Premium allocation approach (PAA) revenue recognition is a distinct measurement simplification and has its own period-allocation rule. The object here is the service measure and its consequences for releasing CSM under the general measurement model, variable fee approach, qualifying reinsurance contracts held, and the special requirements for investment contracts with discretionary participation features.
Insurance Coverage Services — Scientific Subject and Accounting Foundations
Insurance service as continuing performance
A contract may transfer insurance risk at inception, but the insurer performs the coverage obligation by remaining ready to investigate and pay valid claims if an insured event occurs during the covered interval. The existence of insured risk and the delivery of service are related through contractual performance. The service continues even where the realized claim count is zero. The 2022 annuity coverage decision makes the connection concrete: for a group promising periodic payments contingent on survival, the service concerns the coverage afforded for survival during the relevant period, evaluated with regard to the amounts payable when that insured event occurs.
This distinction is foundational. For a term life policy with a fixed death benefit, a policyholder protected during a claim-free month still received the benefit of available coverage. A count of paid death claims measures realized loss payments, while coverage units must reflect the service provided to all contracts in the relevant group. Likewise, technical reserves are estimates of an insurer's liability and the risk adjustment expresses compensation required for bearing nonfinancial risk. Neither figure is automatically a measure of the policyholder's quantity of insurance coverage. These differences are developed in the actuarial CSM analysis of Yousuf et al. (2021) and in the Canadian coverage-units educational note.
The service focus also explains why the expected financial surplus of a contract group can differ from the CSM recognized in a particular year. Expected profit reflects many current estimates and constraints; recognized service margin is a component of current-period insurance revenue arising from prior deferral and subsequent adjustments. The IFRS Foundation's overview of insurance financial performance describes service results separately from insurance finance income and expenses. Accordingly, a profit-recognition study must identify the relevant component rather than equate the release of CSM with a carrier's entire net income.
From IFRS 4 diversity to IFRS 17 service-based accounting
Under the predecessor IFRS 4, national insurance accounting practices persisted in different forms. The international standard-setting project sought greater consistency in valuation and performance reporting by requiring a common measurement architecture. The IASB's IFRS 17 project explanation links the new regime to current estimates of future cash flows, explicit unearned profit, and recognition of profit as service is provided. Coverage units are the temporal allocation device that carries these premises into CSM recognition. They do not supply a substitute for actuarial pricing or commercial management's preferred earnings target.
The requirement is asymmetric in its treatment of favorable and unfavorable expected results. A profitable group ordinarily contains a positive CSM representing unearned profit, while an onerous group requires immediate recognition of a loss under IFRS 17's applicable requirements rather than the establishment of a negative unearned-profit balance. The peer-reviewed life-insurance CSM working-party paper analyzes this structure in depth. Coverage-unit allocations therefore describe the release of a positive remaining CSM; their arithmetic cannot be used to spread an initially recognized onerous loss as if it were future service profit.
What the service measure is not
Premiums received are financing and contractual-consideration flows. They need not coincide with the timing of insurance service; some are paid at inception, others periodically, and some contracts combine insurance and substantial investment components. Expected claims describe future insurer outflows conditional on risk events and can be sharply concentrated in a late duration even where substantial protective service is supplied earlier. The release of the risk adjustment for nonfinancial risk follows its own measurement logic. Expected claims, premiums, projected fulfillment cash flows, surrender values, sum insured, and risk exposure can each be relevant evidence for some products, but none is a universal coverage-unit definition.
A useful discriminating question is: what amount could the policyholder legitimately receive on the relevant insured event, and during what interval is the insurer obliged to make that protection available? For disability cover, salary-indexed periodic benefits may be the relevant service measure. For a fixed death-benefit contract, maximum contractual cover can be an appropriate proxy. For mixed benefits, one must account for each separately identifiable service and avoid suppressing a material rider simply because its amount is smaller or expressed in another unit. This point is substantiated by Yousuf et al.'s coverage-unit discussion and the June 2022 Committee's annuity analysis.
Unit of account and the aggregation problem
IFRS 17 measures and presents insurance contracts in groups, subject to prescribed grouping requirements. An individual contract provides the granular information from which service quantities and coverage durations are assembled; the CSM release is then determined at group level. A group can contain contracts issued at different dates within an eligible cohort, multiple lives, distinct amounts of coverage and differing probabilities of lapse. An algorithm that simply recognizes one equal margin increment for each contract in force is sufficient only where that count genuinely approximates the relative services provided. For mixed benefits or unequal face amounts, unweighted contract counts can erase economically important distinctions.
The consequence is methodological: an actuarial projection can be accurate for survival probabilities while the accounting service measure is still defective. Validity requires two independent connections, first from the contract promise to the chosen benefit measure and then from the projected duration and service measure to each group's coverage units. Grouping and service measurement cannot be treated as interchangeable steps. The IFRS 17 aggregation and measurement overview supplies the overarching structure, while Palmborg, Lindholm, and Lindskog (2021) examine the relationship between financial position, insurance service and the resulting performance measures.
Service Quantity and Duration — Authoritative Requirements and Financial Concepts
The B119 allocation sequence
Paragraph B119 establishes a three-part calculation: identify coverage units representing insurance contract services in the group; allocate the CSM at the end of the period before its current-service release across units supplied in the current period and expected for the future; recognize in profit or loss the amount allocated to units supplied during the period. The official AASB 17 text of B119 directly supports all three steps. Paragraphs 44(e), 45(e), and 66(e) connect the allocation to the respective CSM adjustments for insurance contracts without direct participation, those with direct participation, and reinsurance held.
Define an allocation reporting date indexed by (t). Let (M_t) be the nonnegative CSM immediately before release for service supplied in period (t), after the other applicable subsequent-measurement changes. Let (U_t) be coverage units supplied during the period and (U_{t+s\mid t}) the number expected in future period (t+s), evaluated on the current reporting-date information set. When the total relevant coverage units are positive, the release is
[ R_t=M_t\frac{U_t}{U_t+\sum_{s\geq1}U_{t+s\mid t}}, \qquad M_t^{\mathrm{after}}=M_t-R_t. ]
The expression implements equal allocation per unit within a defined group and unit measure. It is an analytical representation of paragraph B119 rather than a separate accounting rule. Units can be scaled in an internally consistent manner without changing the ratio. For example, expressing all face-amount-based units in dollars rather than thousands of dollars multiplies both numerator and denominator by the same factor. Changing the relative weights of different periods, benefits, or services, however, changes the allocation and requires a substantive justification.
If no current service units exist but future service remains, the equation allocates none to current service. If all remaining units are supplied in the current period, it releases the full pre-release CSM for that group, assuming no separate constraint. Zero denominators, contract modifications, coverage cessation and derecognition require investigation under the underlying requirements rather than forced computation. A negative CSM must not be introduced into this expression as a substitute for onerous-group accounting.
Benefit quantity and service period are two distinct dimensions
The number of units is determined by considering both quantity of benefits and expected coverage period. A nominal benefit amount measures the size of a contractual promise in a covered interval; expected duration measures how long contracts continue providing service. The two can interact. If a life policy provides a fixed benefit of 100,000 currency units, there is no automatic assumption that it provides 100,000 units every future year with certainty. Expected deaths, lapses, benefit reductions, policy options and coverage terminations affect which future services are expected to be supplied.
For a homogeneous group, an illustrative expected-service proxy may be written as (U_t=N_0p_t b_t), where (N_0) is the original covered contract count, (p_t) is the probability a contract remains eligible for coverage during the chosen service interval, and (b_t) measures the quantity of covered benefit supplied while eligible. The probability and timing convention must align with when insurance service is transferred. This is an actuarial approximation, not the wording of IFRS 17, and it need not apply when contract-level projection data are available or multiple benefits exist. As emphasized in the Canadian professional guidance, the determination involves actuarial judgment about quantities, termination and product structure rather than a single mandated formula.
This separation is especially important for an annuity. An immediate life annuity provides insurance coverage against survival in each interval during which periodic payments are contingent on the annuitant being alive. A stream of payments that becomes payable across future intervals measures future service, not all service transferred today. The final 2022 IFRS Interpretations Committee decision articulates the service interpretation and makes clear that IFRS 17 leaves the entity to select a method that actually satisfies the B119 principle.
Claims incurred versus claims payable if the event occurs
Consider a hypothetical group with 1,000 lives, each insured for a fixed death benefit, and with a very low annual death probability. The group provides protection to substantially all in-force policyholders during the year, although few death claims are expected to be paid. A service measure derived solely from expected claim cash flows would be driven by the probability of the insured event. A coverage benefit measure may instead be based on the claim amount for which the insurer stands ready when a valid event occurs, adjusted for contracts expected to remain in force. The distinction is that coverage measures the economic protection available, while expected claims measure estimated cost of fulfilling that promise.
The IFRS Interpretations Committee's discussion of benefit quantity refers to periods in which the insurer must pay valid claims and to the claim amount if the event happens. The June 2022 annuity agenda decision addresses this explicitly. These principles do not mean that maximum contractual limits are always a superior proxy: for some variable indemnity benefits, benefit quantity may need a more representative contractual measure. The ultimate test remains whether the pattern depicts services provided across time.
Expected coverage period versus settlement period
Insurance claims can be filed, investigated and settled years after the insured event. Some of these obligations become part of the liability for incurred claims after the insured event. The period during which an insurer remains obliged to settle already-incurred claims need not mean that the entity is still providing the original insurance coverage. The contractual distinction between remaining coverage and incurred claims underlies the IFRS 17 defined terms and model. Treating claim-runoff time as coverage time without identifying a qualifying service would defer CSM beyond its relevant earning period.
The issue becomes complex for long-tail liability products with extended reporting periods, claims-made wordings, reinstatements and contractual renewal options. The proper coverage interval depends on the contract's insured-event trigger and the obligations actually supplied. Financial, legal and actuarial teams must establish those facts before assigning units. It is inadequate to infer coverage duration mechanically from reserve runoff or from the date on which the last claim is finally paid.
The development of the authoritative interpretation
The Transition Resource Group for IFRS 17 (TRG) examined the determination of benefit quantity soon after the standard's issuance. The February 2018 TRG meeting agenda lists determining the quantity of benefits for coverage units as a specific agenda item. The IFRS Foundation's TRG record establishes that these implementation discussions occurred between February 2018 and April 2019. They exposed the need to distinguish compensation for insured events from other services when groups contain complex life and investment contracts.
In January 2019, the IASB tentatively supported expanding the general measurement model's coverage-unit analysis to investment-return services in relevant cases; the January 2019 IASB Update documents the interim proposal. Following consultations, the February 2020 IASB Update records decisions that refined the investment-return service criteria and added disclosures regarding relative service weighting. The June 2020 amendment project record establishes the final amendment stage. The temporal distinction matters: 2019 tentative judgments are historical evidence of standard development, while the amended IFRS 17 text determines the applicable accounting requirements for later reporting periods.
Coverage Unit Assumptions — Recognition, Valuation, and Quantitative Methodology
A contract-to-unit research procedure
A sound coverage-unit methodology starts with a service inventory. The inventory specifies the legally enforceable promises under each product, distinguishes insurance coverage from investment-return or investment-related services where applicable, and identifies each service's commencement and termination. It then defines the observable quantity of promised benefits, the contractual and actuarial factors influencing expected duration, and the aggregation approach that combines the resulting services into group-level units. Each entry should have a verifiable contract clause or product specification, a measurement rule, an assumption source, and a documented reporting-period convention.
After defining benefit quantity, the projection estimates exposure by period. The relevant inputs can include mortality, morbidity, disability incidence and duration, lapse and surrender, contract status, annuity survivorship, benefit escalation, paid-up options, sum-assured reductions, and the expected exercise of discretionary rights. The appropriate inputs vary by product. A term life policy and a whole-life participating contract need not employ the same benefit basis. The Canadian Institute of Actuaries' 2022 note expressly addresses life insurance, annuities, accident and sickness, segregated funds and group insurance, including contracts that combine different forms of coverage.
At the final stage, the reporting-date calculation takes the pre-release CSM and applies the ratio of current to total current-plus-future units. The future units must reflect the current evidence and estimates. This procedure must be embedded in controls for data lineage, contract changes, reconciliations, projection-version tracking, consistency of methods and management signoff. It converts an actuarial projection into an auditable accounting measure without promoting any particular convenience proxy to the status of a universal rule.
Benefit scaling, currency, and dimensional consistency
A method based on benefit amount requires attention to units of measurement. One dollar of lump-sum death cover and one dollar of periodic income paid across multiple months need not describe equivalent quantities of service. A dollar of managed investment assets is an especially different object. To combine distinct service benefits, the entity must select and justify their relative weights so that each contributes appropriately to the period's composite coverage units. A weighted sum is a possible computational implementation, but weights cannot be chosen merely to reproduce a target earnings curve.
Where contracts are denominated in multiple currencies, currency conversion, determination of group denomination and the CSM's presentation need to be handled under the relevant requirements. A translated amount of benefit is not automatically the economically correct service weight. The IFRIC June 2022 discussion of multi-currency groups underscores that a group has one CSM and must apply a single coverage-unit determination method even when the accounting for monetary amounts involves multiple currencies. Exchange-rate effects must be kept conceptually separate from changes in the quantity of services.
The role of actuarial expectations
The expected coverage period is forward-looking and changes when experience changes. For a portfolio with terminations, expected future units are the sum of services anticipated under still-effective contractual promises, allowing for persistence and mortality estimates. The future-unit denominator is therefore dynamic. It may decline faster than initially expected after unexpectedly high mortality or surrenders, or rise relative to prior expectations where contract durations extend or benefit amounts increase.
The current service period requires its own timing convention. The actual dates of cancellation, death, surrender, new cover, rider activation and benefit changes can influence the amount of service supplied within the reporting interval. Monthly or daily exposure methods can better approximate actual time on risk than an annual start-of-year count when such differences are material. Yet the additional detail is only useful if the underlying service definition is sound. The Yousuf et al. (2021) analysis of remeasurement illustrates alternative treatments of midperiod decrements and explains why practical precision must be evaluated against its effect on service representation.
Discounting the service denominator
IFRS 17 requires equal allocation across coverage units, but its standard-setting materials do not prescribe a single universal answer to whether and how the time value of money should be reflected in the unit denominator. The actuarial discussion of discounting coverage units identifies an important judgment, especially for long-duration contracts with escalations and high rates. Discounting future service units changes their relative numerical weight, and therefore alters the time profile of released CSM. It is not itself a new cash-flow valuation, nor does discounting automatically become appropriate merely because fulfillment cash flows are discounted.
For a transparent sensitivity test, assume that the projected service quantities are ((1,1,1)) in periods 1–3 and that the sole exercise is an allocation of a fixed 120-unit CSM with no subsequent remeasurement. Without discounting, the initial service weights are equal and a static allocation assigns 40 units to each period. If future service quantities are discounted from period 1 at 5 percent per period, the relative weights at that initial date are (1), (1/1.05), and (1/1.05^2), which produce an illustrative first-period allocation of approximately 41.97 units rather than 40. This is an ex ante static comparison of two weight vectors; a genuine multi-period accounting projection must roll forward the CSM, update future units at each reporting date, apply relevant accretion and remeasurement, and avoid freezing a denominator after the date on which the standard requires reassessment.
The mathematical lesson is that a nominally equal volume of future coverage can correspond to unequal present allocation weights if discounting is employed. The accounting lesson is that the selection requires a service rationale and consistent implementation. A common discount rate is not evidence that all benefit types should be discounted identically.
Group homogeneity and representational error
Consider two hypothetical contracts in one eligible group. Contract A provides 100,000 units of death coverage; Contract B provides 10,000 units. If both provide identical duration and the selected benefit measure is death-cover amount, one contract-year for A yields ten times the units of one contract-year for B. An unweighted headcount would treat their services as equal. This is a direct representational failure under that chosen measure. On the other hand, if two products have different risks but similar annual coverage benefit quantities, using expected claim cost as a measure could make the higher-risk product appear to supply more coverage merely because its insured event is more likely. The two examples demonstrate that neither headcount nor expected claims is universally a faithful service quantity.
The diagnosis can be formalized by comparing an implemented proxy (\widehat U_t) with a reasoned benchmark (U_t) constructed from contractual benefits. The ratio of releases is driven by (\widehat U_t/\sum\widehat U) rather than (U_t/\sum U). When the two fractions differ, the resulting earnings shift is a measurement-choice effect. The benchmark itself remains judgmental: it must be constructed from the insurance services and contractual evidence relevant to the contracts, rather than inferred from a preferred earnings outcome. This is a methodological diagnostic, not a new named model.
Service Pattern and Profit Recognition — Reproducible Accounting Examples
Experimental conventions and verification
The worked examples below are original transparent arithmetic exercises. They deliberately omit claims experience, changes in fulfillment cash flows, risk-adjustment changes, acquisition cash-flow allocations, currency translation and tax unless the example explicitly includes a stated component. Amounts and portfolios are hypothetical; they are not observations of an insurer. Each exercise tests one dimension of B119 and documents the equal-per-unit allocation identity. The equations concern the release of the CSM component and should not be interpreted as a complete insurance statement of profit or loss. These conventions align with the IFRS 17 B119 allocation requirement.
Experiment A: Uniform service, constant CSM available for allocation
Suppose the group has a pre-release CSM of 120 monetary units at the first reporting date and three equal annual service quantities of 10 units each. Ignore interest and all other adjustments over the exercise. The first reporting date allocates (120\times10/(10+10+10)=40). The remaining CSM is 80. At the second date, the two remaining service quantities are 10 and 10: the release equals (80\times10/20=40). The final release is 40. The sequence is ((40,40,40)), and cumulative release equals the initial 120.
| Period | Current coverage units | Expected later units | Pre-release CSM | Current-period release | Post-release CSM |
|---|---|---|---|---|---|
| 1 | 10 | 20 | 120.00 | 40.00 | 80.00 |
| 2 | 10 | 10 | 80.00 | 40.00 | 40.00 |
| 3 | 10 | 0 | 40.00 | 40.00 | 0.00 |
This is the limiting case in which an equal yearly release is supported by the chosen service measure. It is a result of equal unit supply under fixed assumptions; it is not a default accounting instruction that service must always be allocated linearly.
Experiment B: Declining service amount and nonuniform recognition
Let an initial pre-release CSM equal 280 with the projected unit profile ((100,80,60,40)). Assume no subsequent changes or accretion. The relevant service sum is 280, so one monetary unit of CSM is assigned to each unit of service for this artificial construction. The yearly releases are ((100,80,60,40)). At the first year-end, the remaining CSM is 180; at the second, 100; at the third, 40; and after the fourth it is zero.
| Period | Current units | Future units | Pre-release CSM | Released CSM | Closing CSM |
|---|---|---|---|---|---|
| 1 | 100 | 180 | 280.00 | 100.00 | 180.00 |
| 2 | 80 | 100 | 180.00 | 80.00 | 100.00 |
| 3 | 60 | 40 | 100.00 | 60.00 | 40.00 |
| 4 | 40 | 0 | 40.00 | 40.00 | 0.00 |
A mechanically straight-line recognition over four years would give 70 per year. It differs from the selected benefit-based service profile by (+30,+10,-10,-30) when measured as coverage-based release minus a straight-line release. The sign and size of each difference derive solely from the unit distribution, and therefore clarify the earnings consequences of the service definition. For a mortgage-protection product with decreasing death benefit, such a profile may be plausible; its appropriateness still depends on contract features, number of contracts in force, and the service quantity selected.
Experiment C: Interest accretion changes release amounts without altering units
Assume the group starts with CSM 100, three equal projected coverage units ((1,1,1)), and hypothetical 5 percent annual interest accretion on the CSM before each year's release. There are no other changes. The first pre-release balance is 105; the release is 35; and the closing balance is 70. The second pre-release balance is 73.50; the release is 36.75; and the closing balance is 36.75. The third pre-release balance is 38.5875 and it is fully released.
| Period | Opening CSM | Assumed accretion | Pre-release CSM | Units: current / total | Released CSM | Closing CSM |
|---|---|---|---|---|---|---|
| 1 | 100.0000 | 5.0000 | 105.0000 | 1 / 3 | 35.0000 | 70.0000 |
| 2 | 70.0000 | 3.5000 | 73.5000 | 1 / 2 | 36.7500 | 36.7500 |
| 3 | 36.7500 | 1.8375 | 38.5875 | 1 / 1 | 38.5875 | 0.0000 |
The sum of releases, 110.3375, equals the initial 100 plus cumulative hypothetical interest accretion of 10.3375. This example is designed to separate the unit denominator from movements in the available CSM. It illustrates interest accretion as relevant to the general measurement model under stated assumptions; a VFA group requires its own applicable measurement treatment. It is not evidence that service units themselves earn an interest rate.
Experiment D: Updating expected future coverage
Assume the group initially has CSM 400 and expected units ((100,100,100,100)). After year 1, the release is 100 and the post-release CSM is 300. Assume, solely for an isolated allocation experiment, that the period-2 reporting-date unit forecast changes to 100 units currently supplied, then 60 and 20 in the two remaining periods. Assume the changed forecast has no independently modeled CSM adjustment. The period-2 denominator is 180; the release is (300\times100/180=166.6667), leaving 133.3333. In period 3, the release is 100 on 60 of the remaining 80 units. In period 4, the final 33.3333 is released on 20 units.
| Period | Current units | Remaining expected units | Pre-release CSM | Release | Closing CSM |
|---|---|---|---|---|---|
| 1 | 100 | 300 | 400.0000 | 100.0000 | 300.0000 |
| 2 | 100 | 80 | 300.0000 | 166.6667 | 133.3333 |
| 3 | 60 | 20 | 133.3333 | 100.0000 | 33.3333 |
| 4 | 20 | 0 | 33.3333 | 33.3333 | 0.0000 |
Rounding in the table is for display; unrounded fractions reconcile. The exercise shows that revised service forecasts can accelerate subsequent recognition without retroactively changing already earned period-1 service. In live IFRS 17 accounting, the events causing changed coverage expectations can also change fulfillment cash flows and the CSM itself, and any such consequences must be recorded independently. The 2021 actuarial examination of future-unit reassessment explains this important interaction.
Experiment E: Annuity service based on survival in each period
Consider an invented group of immediate life annuities paying an annual 12,000 benefit to each annuitant who survives to the corresponding payment date, with projected eligible survivors of 90, 80 and 70 in the three periods. Assume the coverage service is the current period's survival-contingent annual payment; no additional investment service is identified for this simplified illustration. Multiplying eligible survivor counts by 12,000 generates benefit-proportional units ((1,080,000,960,000,840,000)), equivalent in relative proportions to ((90,80,70)). With an available CSM of 240 and no further changes, the releases become ((90,80,70)).
A rival measure that assigns the current period the present value of all later possible annuity payments would potentially represent services of later years as if supplied now. The Committee's 2022 coverage decision instead tests methods against the survival protection actually furnished in the period. This illustration captures its key temporal discipline; it does not establish a universally mandatory annuity-unit formula and does not substitute for determining what other qualifying services the actual contract provides.
Experiment F: Combining insurance and qualifying investment services
Suppose a hypothetical contract group supplies an insurance coverage quantity of 100 in each of three years and also qualifies for a separately identified investment-return service, with illustrative unit scores of ((0,100,200)). The scores are arbitrary normalized representations of benefit quantity, not raw asset values or investment performance, and are used only after the eligibility and weighting assumptions have been made explicit. Let the entity's supported relative weighting assign coefficient 0.5 to the investment-service scores. The composite units are then ((100,150,200)), totaling 450. With CSM 240 and no remeasurement, releases are approximately ((53.3333,80,106.6667)). Insurance-only units ((100,100,100)), by contrast, would release ((80,80,80)).
| Period | Insurance-service score | Investment-service score | Weighted composite units | Release with both services | Insurance-only comparator |
|---|---|---|---|---|---|
| 1 | 100 | 0 | 100 | 53.3333 | 80.0000 |
| 2 | 100 | 100 | 150 | 80.0000 | 80.0000 |
| 3 | 100 | 200 | 200 | 106.6667 | 80.0000 |
| Total | 300 | 300 | 450 | 240.0000 | 240.0000 |
This is a deliberately simplified sensitivity demonstration. IFRS 17 does not prescribe the 0.5 weight, the investment-score construction, or any method that forces later recognition; the actual measure must reflect relative benefits and service transfer. The IASB's February 2020 decisions recognize the need for judgment and disclosure regarding the relative weighting of insurance and investment services. The example's scientific function is to make visible the consequence of a declared weight and to show why an undocumented weight threatens comparability.
Arithmetic and conceptual integrity tests
Each scenario can be tested through a compact set of accounting invariants. First, each period's release must equal the current proportion of total service units multiplied by the relevant pre-release CSM. Second, the closing CSM must reconcile to the pre-release CSM less the period's release. Third, when all modeled service units have been delivered and no other balance is retained, the final CSM must be zero. Fourth, total releases must equal the initial CSM plus all explicitly modeled subsequent additions less any other separately recorded deductions, subject to the exact modeled sequence. Fifth, all quantities used to form a composite service measure must be dimensionally comparable after justified normalization.
These tests detect arithmetic defects and some inconsistencies in system implementation. They do not prove that the coverage-unit definition depicts the genuine service promised under a contract. Scientific validation requires both reproducible calculation and substantive review of the underlying measurement premise.
Variable Service Profiles — Product Application and Source-Based Limits
Level term assurance and decreasing term protection
Level term assurance provides the archetype of a fixed protective amount, subject to the number of in-force policies and the insurer's obligation during each period. With unchanged benefit and no expected termination before contractual expiry, similar covered periods can produce similar service quantities. A decreasing term policy, by contrast, may have an outstanding cover amount linked to a declining debt; a service measure that reflects the reducing claim amount can produce a declining unit curve even if contract counts remain constant. The Canadian coverage-units note and Yousuf et al.'s life-product examples explain why benefit amount and expected period must be identified jointly.
The fact that risk of death increases with age does not by itself establish an increasing quantity of death-cover service. Frequency of insured events and contractual benefit amount are different dimensions. A mortality assumption enters the expected duration of coverage and anticipated persistence of contracts, and can also affect other CSM measurement components. It need not mechanically transform the available face amount into an expected loss measure for B119 purposes.
Whole life, paid-up status, and participating policies
Whole-life assurance introduces uncertain survival duration, possible premium cessation, reduced paid-up benefits, and sometimes bonuses or investment-related returns. A policyholder can remain insured after premiums stop, so coverage cannot be presumed to terminate at the last premium receipt. A paid-up option can reduce death benefits while extending the continuation of service. The actuarial paper's paid-up coverage illustration provides a direct example of why an expected benefit-quantity path must incorporate status changes.
Participation features require a second question: does the contract provide investment-related service under the direct participation criteria, or qualifying investment-return service in the general model? The source of investment returns, policyholder rights, underlying items and contract structure determine the answer. Investment-service identification can extend the relevant service pattern beyond the most visible period of risk coverage, while the end of investment service remains constrained by the policyholders' rights and payments under paragraph B119A. The operative investment-service criteria distinguish these questions from arbitrary investment-asset allocation.
Immediate and deferred annuities
Immediate annuities supply recurring coverage against survival. The accounting unit must represent survival-contingent cover provided during the period, considering both period benefit and expected survival duration. The IFRIC's finalized annuity interpretation shows that selecting a technically plausible benefit amount is insufficient if it allocates future service to the present. That decision resolves the analytical test for its described fact pattern without prescribing a complete list of acceptable methodologies for all annuity designs.
Deferred annuities add an accumulation phase that can involve withdrawal rights, credited investment returns and later insurance coverage during payout. An entity must examine whether the accumulation phase provides qualifying investment-return or investment-related service. A deferred contract with a guaranteed conversion option, death protection, surrender rights or non-guaranteed payout rates may differ from another contract marketed under the same generic product label. The IASB's 2020 investment-service decisions and Canadian annuity-specific guidance support case-specific service analysis rather than an automatic start date or standardized asset-based proxy.
Health, disability, and long-term care
Health and disability contracts can have multiple service components: reimbursement of medical expenditures, fixed hospital benefits, income replacement, waiver of premiums, morbidity-triggered annuities and support during extended disability. A fixed sum insured is a defensible benefit measure for some products, while a salary-linked indemnity or daily reimbursement cap may better represent others. Expected remaining service durations depend on morbidity, continuation, benefit periods, renewals and product definitions. A coverage-unit model must preserve the difference between the amount of benefit accessible after the relevant insured event and the expected present cost of that event.
Where long-term care or disability cover continues during claim payment after the insured event, careful contract analysis is necessary. A claimant may be receiving performance of an incurred claim, continued coverage for additional insured events, an investment service, or a combination. The correct duration follows the contractual promise and the precise IFRS 17 definitions, not a general expectation that all long claims imply a longer coverage period.
General insurance and shorter-duration business
Property and casualty groups often provide a relatively straightforward coverage service measured by insured amount or by an appropriate measure of protection available during the coverage period. The PAA may be used where its eligibility requirements are met; that approach applies a different rule for allocating expected premium receipts to periods. Under AASB 17 paragraph B126, the general PAA basis is passage of time, with a different basis when expected risk release differs significantly from passage of time. This is not the B119 CSM release calculation and must not be conflated with it.
For longer-duration P&C groups under the general model, coverage units continue to matter. Insurance service can be highly nonuniform for seasonal catastrophe cover, construction stages, or fluctuating commercial exposure. A purely elapsed-time denominator may fail to depict the differing levels of cover where the contract benefits change substantially during a period. Any chosen proxy must be checked against the benefit actually provided and the remaining contractual service period.
Reinsurance contracts held
IFRS 17 treats reinsurance contracts held separately from underlying insurance contracts issued. Under the general model for reinsurance held, CSM recognition reflects reinsurance services received, with the relevant standard adjustments, rather than simply mirroring the direct insurance CSM release. Paragraph 66(e), referenced by B119's application text, connects reinsurance-held accounting to service-based CSM allocation. Quota-share, excess-of-loss and multi-year arrangements require close analysis of reinsured benefits and coverage period to identify the transferred service. A premium ceded or an expected recovery amount is not automatically the same thing as the quantity of reinsurance service received.
This dossier does not calculate reinsurance loss-recovery components or the asymmetric recognition of certain recoveries for onerous underlying groups. Those accounting mechanisms are covered in their own domain. They matter here because one must not assume that the direct and ceded service patterns will be synchronized simply by applying the same unit ratio to both balances.
Investment contracts with discretionary participation features
IFRS 17 also contains specified modifications for investment contracts with discretionary participation features in its scope. The requirements modify CSM allocation so that the margin is recognized over the group's duration in a systematic way that reflects transfer of investment services. The text of the special investment-contract provision supports that distinction. Contract classification is therefore prior to selecting a profit-allocation denominator. Applying insurance-coverage units to a contract whose operative accounting service is investment service would misidentify the activity the standard asks the financial statements to portray.
Cross-product lessons
Four product properties repeatedly determine service measurement: the event protected, the conditional amount the insured or beneficiary may validly receive, the duration of enforceable provision of that protection, and the existence and timing of additional investment services. Product names are secondary. The same marketing label can describe different promises across jurisdictions, and similar benefit structures can appear under different distribution and statutory regimes. This is why the research conclusion rests on rights and obligations rather than on a fixed product taxonomy.
Disclosure and Estimation Sensitivity — Comparative Interpretation and Analytical Synthesis
What users of financial statements can observe
External users do not ordinarily see a contract-by-contract spreadsheet of coverage units. They see insurance revenue, insurance service results, CSM reconciliations and disclosures concerning judgments and the expected timing of future CSM recognition. Paragraph 109 requires quantitative information in time bands on expected CSM recognition; paragraph 117 addresses significant judgments and estimation methods. The AASB 17 disclosure requirements make these requirements visible in the official adapted text. The IASB's February 2020 deliberations additionally document the intended disclosure of relative weighting between insurance coverage and investment services.
A reviewer needs a service-based explanation of the CSM release rather than a single label such as 'straight line' or 'sum assured.' A useful account states what services are included, how benefit quantities are measured, why the aggregation of different services is reasonable, which assumptions determine future duration, how economic and demographic changes affect the denominator, and whether the method changed. These are not substitutes for compulsory disclosures; they are practical implications of the standard's demand for material judgment transparency.
Financial comparability and discretion
IFRS 17 creates a common architecture for insurance financial reporting, but principle-based measurement still permits differing reasonable implementations. In the same product segment, two insurers can choose distinct yet supportable service-benefit measures or different approaches to the time value of money, resulting in different annual CSM release profiles. The Yousuf et al. (2021) analysis identifies several such judgments and their potential materiality. The Palmborg, Lindholm, and Lindskog article supplies a complementary formal examination of financial position and performance under the standard.
Comparability should therefore be tested at three levels. Legal comparability concerns whether the contracts offer similar promises. Measurement comparability concerns whether the entities choose comparable service quantities and durations, normalize composite services consistently, and update assumptions on similar grounds. Reporting comparability concerns whether the effects of different choices can be identified from reported CSM, revenue and disclosure. Equal reported CSM release rates do not prove equal service and unequal release rates do not prove that either entity misapplied IFRS 17. The underlying contract and accounting methods must be examined.
Disclosure-driven analytical reconstruction
An investor or researcher can construct a limited bridge from observed CSM movements to service release. The starting variable is the amount of CSM before current-period service recognition after other required adjustments. Reported current service CSM release reveals a ratio (R_t/M_t), if both amounts can be identified accurately. This ratio is mathematically equivalent to current coverage units divided by current-plus-expected-future units under the specified allocation convention. The bridge does not identify the absolute quantity of units because multiplication of all units by a common constant leaves the ratio unchanged. Nor can it identify the distribution across distinct services without further information.
This identification limitation is fundamental. For example, observing a 25 percent release ratio is consistent with 25 current and 75 remaining units, or 250 current and 750 remaining units, or a composite construction that weights investment-service units differently. It can confirm the applied fraction while leaving the substantive service measure unidentified. Relevant accounting-policy disclosure and contract information are therefore necessary for scientific comparison, even when the arithmetic roll-forward reconciles perfectly.
Sensitivity of the released amount to estimates
Take the basic formula (R=M U/(U+V)), where (U) is current units and (V) is future units. Holding (M) and (U) fixed, the partial derivative with respect to future units is
[ \frac{\partial R}{\partial V}=-\frac{MU}{(U+V)^2}. ]
For a positive CSM and current service, a higher expectation of future units reduces current CSM release, while a reduction of expected future units increases it. Holding (M) and (V) fixed,
[ \frac{\partial R}{\partial U}=\frac{MV}{(U+V)^2}. ]
Increased current service raises current release when future service remains positive. These results follow directly from the B119 ratio and apply only when other determinants are held fixed. In actual reporting, the assumptions that alter units may also change projected cash flows and thereby the pre-release CSM; that joint movement can offset or amplify the pure denominator effect. Treating a single-parameter derivative as a causal estimate of full profit is therefore an analytical error.
The elasticity of release with respect to future units is (-V/(U+V)); the elasticity with respect to current units is (V/(U+V)). Their magnitudes illustrate an important reporting fact: when much service remains, a proportionate revision to the denominator can materially affect current release; near final runoff, the range of future service expectations is smaller but the amount of CSM left to allocate may also have changed. These are algebraic implications of the accounting ratio, offered as an original explanatory synthesis rather than a purported new actuarial doctrine.
Risk of earnings management and boundaries of judgment
Because the model requires judgment about the amount and timing of service, management may have incentives to select methods that produce smoother, earlier or later reported margins. Professional judgment is legitimate when it accurately captures contractual services; its existence creates a corresponding need for evidence, controls and independent review. To assess an apparent accounting-policy change, one should ask whether an actual product feature, contract mix or better evidence explains the change, whether similar contracts are treated consistently, and whether the financial statement disclosures state why the method changed.
An effective control system should restrict arbitrary revision of service weights, preserve earlier projections, record each assumption change with its supporting experience data, and reconcile outputs to contract registers and CSM roll-forwards. A controlled comparison of alternative reasonable methods can be useful to evaluate materiality. However, selecting whichever method meets a commercial earnings budget would invert the purpose of B119, converting service representation into earnings design.
International reporting context
IFRS 17 is issued by the IASB as an international accounting standard, but local adoption and implementation take place in individual legal jurisdictions. A nationally adopted standard, such as Australia's AASB 17 Insurance Contracts, is an authoritative reference for the period and entities to which that instrument applies. It is also a useful accessible reproduction of much IFRS 17 wording. The Australian web version consulted here is operative for reporting periods beginning on or after July 1, 2026 and before January 1, 2027; this particular local version must not be described as universal law in every jurisdiction or as the original IASB publication. IFRS 17's internationally stated effective date remains January 1, 2023 for applicable annual periods, subject to jurisdictional adoption and entity eligibility.
National professional actuarial bodies can inform judgment within that framework without altering the authoritative standard. The Canadian Institute of Actuaries note is guidance particularly useful for life and health insurance practitioners; the Institute and Faculty of Actuaries working-party research demonstrates choices arising in life insurance. Neither substitutes for a specific entity's contractual assessment, local reporting obligations or audited financial statements. U.S. GAAP insurance reporting also has its own requirements; the IFRS 17 coverage-unit rule cannot simply be projected onto U.S. GAAP long-duration insurance accounting without a separately verified standards comparison.
Source criticality and historical interpretation
The evidence base has three distinct levels. At the normative level are the operative standard and final interpretation decisions. At the institutional explanatory level are TRG discussions, IASB deliberations and professional educational materials that illuminate drafting history and application. At the analytical level are peer-reviewed actuarial studies and the reproducible calculations in this dossier. The levels should be connected but not collapsed: an illustrative numerical result demonstrates the consequences of assumptions, while only an authoritative requirement can establish what reporting rule applies.
The history of investment-service changes is a cautionary example. The January 2019 Board position was tentative and narrower than the final criteria. The February 2020 decisions revised the framework, and the 2020 IFRS 17 amendments were issued afterward. A comparative analysis that cites the earliest proposal as the current test for investment-return service is historically and substantively incorrect.
The annuity interpretation supplies another distinction. The June 2022 Interpretations Committee update records the Committee's finalization and referral for IASB non-objection; the completed project record confirms that the IASB did not object in July 2022. The final interpretive status is consequently established more reliably by the completed project than by a March 2022 tentative announcement.
Analytical synthesis: how service becomes reported profit
The research supports an integrated sequence. A contract determines the insurer's promises; those promises determine services; services define measurable benefit quantities during covered intervals; actuarial assumptions determine expected continuation of those quantities; benefit and duration projections form group coverage units; and the current-to-total ratio allocates the remaining CSM. The resulting recognized amount participates in the insurance service result, alongside other accounting elements that must be kept analytically distinct. Each arrow in that sequence requires its own evidence. An error early in the sequence cannot be repaired merely by perfect arithmetic at the end.
The strongest comparative claim is therefore conditional: coverage-unit accounting improves the conceptual connection between service supplied and profit recognized when the unit measure faithfully depicts the contractual services, and it provides a verifiable calculation structure for that allocation. Its practical informational value depends on the quality of product classification, data, assumptions, weighting and disclosure. The framework's flexibility accommodates genuinely different contracts, while documentation makes the exercise of judgment scientifically examinable. This conclusion follows from the standard, the official interpretive record, actuarial literature and the six calculations; it does not require naming a new theory.
Research limitations and open questions
This monograph examines a targeted corpus of official standards, interpretation history, professional guidance and peer-reviewed research. It does not estimate insurer-specific earnings from proprietary data, audit a selected carrier's live financial systems, establish a statistical industry average, or exhaust all legal adoption variations. Illustrative computations isolate allocation effects and do not model complete CSM remeasurement. AASB 17 is referenced with its clearly stated local operative period. The scholarly journal materials are used according to the access actually available: the Cambridge article was accessible in full HTML, while the Palmborg–Lindholm–Lindskog study's publisher abstract and visible snippets support only the attributed general discussion, not its full detailed empirical or mathematical claims.
Further research can examine public-company disclosures across reporting jurisdictions, quantify the dispersion of coverage-unit assumptions by product and group, construct a replicable comparison of discounted versus undiscounted service weights, and study the interaction of demographic experience and method changes on CSM timing. A particularly important empirical design would hold the contract-level benefit structure and the pre-release CSM fixed while comparing documented methods; this would separate service-measure effects from concurrent changes in fulfillment cash flows. Such research would require a carefully specified sample, dated accounting policies, source financial statements, and replicable microdata or sufficiently detailed disclosures.
Conclusion
Insurance service coverage units govern the temporal allocation of the contractual service margin under IFRS 17 by connecting recognized unearned profit to the insurance contract services provided during the period. The accounting process begins with identification of promises, continues through determination of benefit quantity and expected service duration, and culminates in allocation of the pre-release group CSM between current and future units. The governing requirement is paragraph B119; investment-related and investment-return services and particular reinsurance or investment-contract cases require their additional prescribed analyses.
Four conclusions follow. First, coverage is a continuing contractual service that can be delivered without a claim payment; claim incidence and expected claim cost cannot stand as universal substitutes for service quantity. Second, both the magnitude of benefits and expected duration matter, and mixed benefits require a defensible relative weighting. Third, changes in the projected service profile can redistribute CSM release across periods independently of the effects that those same circumstances may have on CSM remeasurement. Fourth, consistent documentation and meaningful disclosure are the conditions under which a flexible, principle-based allocation rule becomes comparable and auditable.
The worked exercises establish the arithmetic of those consequences and expose the separate influence of service pattern, accretion, estimate revisions and investment-service weighting. They are original analytical illustrations of an existing accounting framework. The remaining scientific task is empirical: to evaluate how well insurers' implemented benefit measures correspond to the services promised by real contracts and how much observable differences in reported earnings arise from measurement discretion rather than product economics. Within its defined scope, the dossier establishes a systematic, source-verified account of the service-to-profit connection required for rigorous international insurance accounting research.
References
Australian Accounting Standards Board. (2022, as compiled for the operative period July 1–December 31, 2026). AASB 17 Insurance Contracts. Australian Accounting Standards Board. https://standards.aasb.gov.au/aasb-17-dec-2022-0
Canadian Institute of Actuaries, Committee on Life Insurance Financial Reporting. (2022, December 18). Educational Note: IFRS 17 Coverage Units for Life and Health Insurance Contracts (Publication 222175). Canadian Institute of Actuaries. https://www.cia-ica.ca/publications/222175e/
IFRS Foundation. (2018, February 6). IFRS 17 Transition Resource Group: February 2018 meeting agenda. IFRS Foundation. https://www.ifrs.org/news-and-events/calendar/2018/february/ifrs-17-transition-resource-group/
IFRS Foundation. (2018–2019). Transition Resource Group for IFRS 17 Insurance Contracts: Mandate, meetings and completion. IFRS Foundation. https://www.ifrs.org/groups/transition-resource-group-for-insurance-contracts/
International Accounting Standards Board. (2019, January). IASB Update: Recognition of the Contractual Service Margin in Profit or Loss in the General Model. IFRS Foundation. https://www.ifrs.org/news-and-events/updates/iasb/2019/iasb-update-january-2019/
International Accounting Standards Board. (2020, February). IASB Update: Contractual Service Margin Attributable to Investment Services. IFRS Foundation. https://www.ifrs.org/news-and-events/updates/iasb/2020/iasb-update-february-2020/
International Accounting Standards Board. (2020, June). Amendments to IFRS 17 Insurance Contracts: Completed Project. IFRS Foundation. https://www.ifrs.org/projects/completed-projects/2020/amendments-to-ifrs-17/
International Accounting Standards Board. (2017, amended 2020 and subsequently). IFRS 17 Insurance Contracts: Issued Standard and Supporting Overview. IFRS Foundation. https://www.ifrs.org/issued-standards/list-of-standards/ifrs-17-insurance-contracts/
IFRS Interpretations Committee. (2022, June). IFRIC Update: Transfer of Insurance Coverage under a Group of Annuity Contracts (IFRS 17), and Foreign-Currency Aspects of Insurance Contracts. IFRS Foundation. https://www.ifrs.org/news-and-events/updates/ifric/2022/ifric-update-june-2022/
IFRS Interpretations Committee. (2022, June; IASB non-objection July 2022). Transfer of Insurance Coverage under a Group of Annuity Contracts (IFRS 17): Completed Project. IFRS Foundation. https://www.ifrs.org/projects/completed-projects/2022/profit-recognition-for-annuity-contracts-ifrs-17/
Palmborg, L., Lindholm, M., & Lindskog, F. (2021). Financial position and performance in IFRS 17. Scandinavian Actuarial Journal, 2021(3), 171–197. https://doi.org/10.1080/03461238.2020.1823464
Yousuf, W., Stansfield, J., Malde, K., Mirin, N., Walton, R., Thorpe, B., Thorpe, J., Iftode, C., Tan, L., Dyble, R., Pelsser, A., Ghosh, A., Qin, W., Berry, T., & Er, C. (2021). The IFRS 17 contractual service margin: A life insurance perspective. British Actuarial Journal, 26, e2, 1–105. https://doi.org/10.1017/S1357321721000015