RESEARCH DOSSIER
IFRS 17 Risk Adjustment and Accounting Disclosure
Abstract. The risk adjustment for nonfinancial risk is a distinct element of the measurement of insurance contracts under IFRS 17. It expresses the compensation an insurer requires for bearing uncertainty in the amount and timing of cash flows arising from nonfinancial risk. Its reported amount influences fulfillment cash flows, the contractual service margin, the measurement of liabilities for remaining coverage and incurred claims, subsequent profit emergence, and the explanatory content of financial statements. The apparent simplicity of a single monetary amount conceals consequential choices involving the risk distribution, diversification, risk tolerance, measurement technique, aggregation, and allocation of changes between insurance service and insurance finance results. This monograph examines the accounting meaning, standards-based scope, quantitative estimation, subsequent movement, and disclosure of that adjustment. Its research method combines close reading of IFRS 17 and its adopted text, analysis of actuarial educational guidance, interpretation of first- and third-year implementation evidence, and reproducible illustrative calculations. It distinguishes risk adjustment from discounted expected cash flows, unearned profit, financial risk, capital requirements, the Solvency II risk margin, and reinsurance counterparty default. It examines the disclosure of methods and equivalent confidence levels under paragraph 119, the role of accounting-policy disclosures and reconciliations, and the possibility that numerically comparable confidence levels describe economically different risk-bearing arrangements. The evidence shows material heterogeneity in methods, disclosure granularity, and diversification practices across reporting entities. Comparative empirical studies identify improvements in reporting structure together with persistent limits to measurement comparability. The central analytical conclusion is that a risk adjustment becomes financially interpretable through a linked explanation of its amount, technique, risk perimeter, diversification, uncertainty, and period movement. Equivalent confidence-level disclosure is essential to IFRS 17 compliance but is only one component of substantive comparability. The discussion is international in scope, avoids treating samples as universal populations, and makes no claim to establish a new accounting theory.
Keywords: IFRS 17; risk adjustment for nonfinancial risk; insurance contracts; actuarial valuation; fulfillment cash flows; contractual service margin; liability for remaining coverage; liability for incurred claims; equivalent confidence level; value at risk; cost of capital; reinsurance held; financial statement disclosure; insurance service result; insurance finance income and expenses; Solvency II; international accounting comparability.
Introduction
The International Accounting Standards Board developed IFRS 17 to provide a coherent accounting framework for insurance contracts across entities and reporting periods. The IFRS 17 standard overview establishes its application to annual reporting periods beginning on or after January 1, 2023. Within that framework, the adjustment for nonfinancial risk is part of the present measurement of the obligations arising from insurance coverage and claim settlement. This liability component has a different economic interpretation from the present value of expected claim payments. It records the compensation required to accept uncertainty, whereas expected cash flows record an expectation over possible outcomes. The distinction operates within the whole contract-accounting system: a liability estimate is not fully understood by knowing its mean alone.
The governing definition lies in IFRS 17 paragraph 37 and its application guidance B86–B92. The European Union's endorsed IFRS 17 text gives the complete legal wording in a publicly accessible adopted version, while the Australian Accounting Standards Board's compiled AASB 17 text provides the corresponding requirements in an operative national accounting standard. Jurisdictional adoption must be distinguished from the IASB's underlying text. The present analysis uses these official texts to identify the reporting requirements and then draws on implementation studies to examine observed practice.
The risk adjustment's significance is evident in three different informational settings. To managers it can represent the reported compensation for exposure to insurance, policyholder-behavior, and operationally specified contractual uncertainties. To investors it contributes to the measured insurance contract liabilities and patterns of reported service performance. To actuaries it requires estimation from underlying outcomes, tail behavior, dependence, and the entity's own risk appetite. These audiences interpret the same number through different analytical questions. A complete disclosure allows the number to be traced back to the uncertainty it represents.
IFRS 17 introduces several distinctions that matter throughout this work. The official IFRS 17 key terms describe fulfillment cash flows, the contractual service margin (CSM), the liability for remaining coverage (LRC), and the liability for incurred claims (LIC). The adjustment for nonfinancial risk is an explicit component of fulfillment cash flows under the general measurement approach. Subsequent changes in that component can have different accounting destinations depending on the service period to which they relate, the measurement approach, and the permitted treatment of financial effects. Reinsurance held requires its own measurement and cannot be reduced to a mechanical deduction from the gross insurance obligation.
The inquiry uses four methods. First, source-critical textual interpretation distinguishes mandatory IFRS requirements from an entity's estimation choice and from educational guidance. Second, quantitative examples illustrate what changes when a confidence level, correlation, capital projection, or reporting choice changes while other inputs remain fixed. Third, observational evidence from EIOPA, KPMG, and academic research tests whether those distinctions explain real disclosures. Fourth, comparative analysis contrasts IFRS 17 with the prudential objective of Solvency II, taking account of legal changes adopted by November 2026 but not yet applicable until January 2027. All numerical examples are hypothetical and auditable; none is described as a reported company's financial result.
Risk Adjustment Accounting Concept
The compensation measured
Paragraph 37 of IFRS 17 states the governing concept: the risk adjustment for nonfinancial risk reflects the compensation the entity requires for bearing uncertainty about the amount and timing of cash flows arising from nonfinancial risk. Application guidance B87 explains the economic indifference underlying that amount: the entity compares fulfillment of a liability with variable outcomes against fulfillment of a liability with fixed cash flows of the same expected present value. The paragraph 37 definition and application guidance therefore give the adjustment a distinct interpretation. It is an entity-specific measurement of compensation, anchored in insurance obligations and the risk the entity actually bears.
This definition immediately clarifies why the adjustment is additional to the probability-weighted estimate. A future claim distribution can have a mean of 100 currency units yet contain outcomes ranging from 50 to 200. Recording an expected outflow of 100 describes the center of that distribution; it does not measure the compensation the insurer would require to accept the distribution instead of a fixed payment of 100. The adjustment supplies that second component. An insurer may calculate both elements using related simulation models, but the reported components answer different questions.
The same paragraph and application guidance preserve a disciplined distinction between uncertainty categories. Insurance risk, lapse risk, and expense risk are examples of nonfinancial risks. Financial risks, including the time value of money and risks arising from financial variables, are reflected in cash-flow estimates or discount rates under the applicable measurement rules. IFRS 17 does not intend a second charge in the risk adjustment for uncertainties already incorporated as financial risk. An actuarial explanation in International Actuarial Note 100 helps distinguish such practices; that educational note is guidance, rather than a source of additional binding accounting requirements.
The amount reported is affected by the entity's required compensation, the extent of diversification benefits recognized, the probability distribution employed, and risk characteristics of the contract portfolio. Paragraph B88 requires attention to the degree of diversification benefit included in determining the compensation required. Paragraph B89 describes characteristics that would normally imply a larger risk adjustment, including less frequent but more severe events, longer-duration uncertainty, and wider probability distributions. These are interpretive features of the standard rather than instructions to select a single universal quantile or cost-of-capital rate.
Distinction from contractual service margin
The contractual service margin represents unearned profit associated with future insurance service, as described by IFRS 17's defined measurement components. For a profitable group measured under the general model at initial recognition, higher fulfillment cash flows—including a higher risk adjustment—ordinarily imply a correspondingly smaller CSM, all else equal. The initial combination is structured so that unearned profit is recognized as the service is provided rather than as an immediate gain at inception. It follows that the adjustment can shift the profile of reported insurance profitability even when the expected cash flows are unchanged.
For a strictly illustrative simplified group, suppose the present value of premiums is 1,000, while the present value of expected claims and directly attributable expenses is 800. If the nonfinancial risk adjustment is 50, the preliminary surplus before the CSM is 150. If it is 80, the surplus is 120. The first group has a larger unearned margin than the second, but each initially recognizes the same net combined amount under the assumed profitable-contract structure. These figures omit acquisition cash-flow allocation, financing timing, nonfinancial-model segmentation, and other real-world accounting complications. They demonstrate the direct connection between the two measurement components and do not replace a contract-level journal-entry analysis.
The distinction is even more important when insurance contracts become onerous. If fulfillment cash flows exceed the available consideration, the general measurement model does not permit a negative CSM to hide the loss. The relevant loss is recognized according to the onerous-contract provisions, including the loss component for the liability for remaining coverage. A large risk adjustment can contribute to an onerous outcome; whether it does so depends on the whole group's measured cash flows and applicable accounting rules, not on the adjustment considered in isolation. The EU-adopted recognition and loss provisions provide the normative basis.
Distinction from regulatory capital and prudential risk margin
IFRS 17 is a financial-reporting standard, while Solvency II is a European prudential valuation and capital regime. Solvency II technical provisions are built from a best estimate and a risk margin defined by regulatory rules, including projected solvency-capital requirements and a prescribed cost-of-capital rate. Its supervisory objective differs from the entity-specific compensation measured under IFRS 17. The Solvency II valuation framework and EIOPA's IFRS 17 implementation report illustrate the distinction.
Consequently, the same internal projection engine may support both calculations without making the outputs conceptually interchangeable. Differences in perimeter, diversification, capital cost, financial assumptions, valuation date, and the treatment of operational and counterparty risks can produce different results. A reconciliation must identify these differences rather than merely assert that one margin is higher. Where a cost-of-capital technique is used in IFRS 17, the adoption of a similarly named technique does not turn the adjustment into the regulatory Solvency II risk margin.
The accounting claim of comparability
The adjustment is an example of a measurement whose vocabulary is standardized more tightly than its numerical recipe. IFRS 17 requires that it have a defined accounting meaning, but it deliberately leaves the estimation technique open, subject to that meaning and the standard's disclosure rules. This creates a specific interpretation problem. Two companies can report adjustments of equal size for portfolios of different uncertainty; two other companies can report different amounts for economically similar portfolios because their required compensation differs. A ratio such as risk adjustment divided by expected claims does not resolve the difference unless the distribution, portfolio horizon, diversification, and accounting boundary are also known.
The useful comparison is therefore multi-dimensional: the monetary amount, the technique, the equivalent confidence level, the associated nonfinancial risk distribution, the included diversification benefit, and the movement in the accounting period. This is an interpretive conclusion drawn from IFRS 17's measurement objective and disclosure architecture, not a proposed new valuation model.
Nonfinancial Risk Exposure
The perimeter of relevant risks
Application guidance B86 of IFRS 17 identifies the risk adjustment's nonfinancial perimeter. Insurance risk includes uncertainty about the insured event's occurrence and severity. Lapse risk includes uncertainty about the continuation of contracts and the corresponding rights and obligations within the contract boundary. Expense risk includes uncertainty about relevant fulfillment expenses. Mortality, longevity, morbidity, catastrophe frequency, litigation severity, reserve development, and policyholder behavior may be prominent according to the nature of the insurer's business. The adopted IFRS 17 application guidance supplies the classification framework.
The perimeter must be interpreted through the contract measurement. An exposure that matters to an insurer as an operating enterprise need not form part of the adjustment for fulfilling an insurance contract. For example, operational risk external to the fulfillment cash flows of insurance contracts belongs to different corporate risk analyses. An impairment of an investment asset is treated through the applicable financial asset accounting. A change in a market discount rate enters the financial measurement of the insurance liability rather than being duplicated as an allowance for nonfinancial risk. The conceptual boundary prevents double counting across distinct financial-statement objects.
The uncertainty being compensated can arise from the amount of a settlement, its timing, or both. A liability expected to pay 100 in ten years with a narrow outcome range has a different risk profile from one expected to pay 100 next year with a small chance of paying 1,000. Even when discounted means are equal, tail characteristics and horizon shape the insurer's compensation. Applying one fixed percentage to all expected liabilities without examining their distributions can obscure the reasons for the risk charge.
Probability distributions and tail behavior
An actuarial model translates contract experience into a distribution of possible future cash flows. It may use frequency-severity methods for nonlife claims, decrement models for life insurance, cohort lapse models, economic and medical trend assumptions, or a stochastic combination of these. The model's distributional form, calibration period, treatment of sparse severe events, and dependence assumptions all influence the proposed amount. The accounting concept does not prescribe a single stochastic model, but the model must be capable of supporting the compensation required for the relevant nonfinancial uncertainty.
A confidence-level technique gives a particularly transparent example. In a simple one-period setting with a loss random variable X, an illustrative risk adjustment at confidence level q can be written as the difference between the q-quantile of X and its expected value:
This formula describes an illustrative quantile loading, not a mandatory universal IFRS 17 formula. It assumes a compatible measurement scale and treats the modeled loss distribution as a sufficient description of the risk. For an exactly normal loss distribution with mean 100 and standard deviation 20, the 75th percentile is approximately 113.49, yielding a loading of 13.49. At the 85th percentile, the loading is approximately 20.73. Changing the percentile alone produces an increase of approximately 7.24, or 53.7 percent of the lower adjustment. Neither scenario changes the expected loss. The sensitivity illustrates why the equivalent confidence level is economically consequential.
The normal distribution is intentionally simple. Long-tailed bodily injury liabilities, catastrophe-exposed commercial insurance, and emerging liability risks may have skewed or heavy-tailed distributions. Their tail quantiles can respond sharply to assumption changes. In such settings the choice of distribution, use of credibility, parameter uncertainty, and treatment of extreme but plausible experience can dominate the choice between two nearby confidence levels.
Diversification and aggregation
Risk diversification is not merely a computational correction. It identifies the portfolio over which the reporting entity recognizes that adverse outcomes can offset one another. The portfolio can be diversified across lines of business, geographical regions, currencies, accident years, and claim types. The resulting reduction in required compensation depends on assumed dependence and on the entity's view of its risk-bearing capacity. The IFRS 17 guidance on diversification and risk characteristics requires an adjustment consistent with the compensation the entity requires, including diversification benefits it considers.
Consider two normal loss components with standard deviations 20 and 30, correlation 0.25, and identical 75 percent quantile-load rules. Separate stand-alone standard deviations sum to 50. The standard deviation of their combined loss is:
At the 75 percent confidence level, adding separate quantile loadings gives approximately 33.72, while applying the quantile loading to the aggregate distribution gives approximately 26.98. The 6.74 difference is a 20 percent reduction from the sum of stand-alone loadings. This result follows from the explicitly assumed normal distribution and correlation; it is not a measured benefit for a real insurer. If correlation increases toward one, the modeled diversification advantage shrinks. If a rare common catastrophe affects both portfolios, a normal-correlation approximation may substantially understate dependence.
Allocation of an aggregate adjustment to contract groups may serve internal measurement and reporting needs. The mathematical allocation rule could be proportional to premiums, stand-alone risk, marginal contribution, or another defensible measure. Under IFRS 17, the resulting group measurements must still be consistent with the overall standard and avoid an internally inconsistent or duplicated adjustment. An allocation that is mechanically stable but economically disconnected from the individual group's risks can create misleading profit patterns.
Data and model uncertainty
Insurance losses are observed with delays, selection effects, reporting changes, and evolving legal or medical environments. A confidence interval estimated from stable claims triangles may underrepresent emerging risk. A pandemic, catastrophe cluster, or unexpected inflation in repair costs can change the loss distribution after historical calibration. Model risk should be treated transparently: an estimated quantile is conditional on modeling assumptions, and a nominal 80 percent quantile is not proof that the model achieves that percentile in future experience.
This distinction has implications for disclosure. A disclosed 80 percent equivalent confidence level is a description of how the reported adjustment relates to an estimated distribution. It does not by itself reveal whether the underlying distribution is calibrated robustly or includes meaningful stress outcomes. An investor evaluating a risk adjustment should ask about model boundaries, diversification, sensitivities, and revisions to assumptions. Paragraph 121 and related disclosure provisions of the adopted insurance reporting requirements explicitly situate reporting within the nature, amount, timing, and uncertainty of cash flows from insurance contracts.
The scope of reinsurance-held uncertainty
The risk adjustment for reinsurance contracts held is governed by the transferred-risk perspective. Under paragraph 64 of IFRS 17, the adjustment reflects the amount of risk transferred by the holder to the reinsurer; it is therefore a different measurement object from the adjustment for contracts issued. Paragraph 63 incorporates the effect of reinsurer nonperformance in estimates of future cash flows rather than placing that credit effect into the nonfinancial risk adjustment. The IFRS 17 reinsurance provisions are decisive for distinguishing transfer of nonfinancial risk from a counterparty credit exposure.
A gross risk adjustment of 15 and a net-retained risk adjustment of 9 may suggest a transferred-risk amount of 6 in an extremely simplified economic picture. They do not authorize the automatic accounting conclusion that the reinsurance-held adjustment must equal exactly 6. The contracts may differ in timing, contract boundaries, reinstatements, deductibles, limits, diversification, and reporting aggregation. The reinsurance-held measurement must be performed on its own contractual facts.
Releases and Measurement Changes
Initial recognition and subsequent remeasurement
The risk adjustment begins as part of the fulfillment cash flows recognized for the relevant insurance contract group. After initial recognition it is remeasured as the entity revises its assessment of nonfinancial risk. Such revisions may reflect updated assumptions, new claims experience, elapsed coverage, settlement progress, changes in dependence, or changes in the compensation the entity requires. This means the adjustment is a moving component of measurement; it should not be treated as a fixed percentage of claims liabilities across the full contract life.
Under the general measurement model, the liability for remaining coverage and liability for incurred claims perform different functions. The LRC represents obligations associated with future coverage and relevant future service; the LIC represents obligations for insured events that have already occurred. Their risk adjustments may be calibrated to different uncertainty. A newly written short-duration portfolio may have substantial uncertainty about the frequency of future claims, while a mature long-tail claims portfolio may have relatively little remaining coverage but considerable uncertainty about judicial outcomes and settlement development. The definitions of LRC and LIC make the distinction explicit.
The accounting destination of a movement depends on why it arises. Changes related to future service can adjust the contractual service margin under the applicable general-measurement requirements, subject to special provisions and any loss component. Changes arising from current or past service are generally reflected in the insurance service result, while the effects of financial risk and the time value of money are considered under the finance presentation provisions. A release of risk as uncertainty resolves is therefore not synonymous with cash claim settlement. Cash settlement extinguishes expected and actual obligations; risk release reflects a reduction in compensation required for remaining uncertainty.
Components of a rollforward
A useful explanatory risk-adjustment rollforward identifies opening adjustment, additions for new business, releases associated with service, changes in assumptions, effects of claims experience and remeasurement, finance-related effects where separately presented, and closing adjustment. IFRS 17 requires reconciliations of the relevant opening and closing insurance contract balances, with disaggregation of the fulfillment components and additional disclosures according to their nature. The IFRS 17 reconciliation provisions in paragraphs 98–101 identify the required information, while paragraph 117(c)(ii) requires a description of how the entity determines the risk adjustment, including its approach to disaggregating the adjustment's changes between service and finance components.
Take an illustrative opening risk adjustment of 50. During the year, new business adds 8, changes in claim and assumption estimates for current/past service reduce it by 5, release of uncertainty reduces it by 15, and the separately modeled finance effect adds 2. The closing amount is:
The numerical identity reconciles an illustrative balance. It does not assert that all five items are individually required captions in every statutory statement, nor that each movement affects profit or loss in the same way. The actual presentation must follow the relevant group and measurement approach, explain whether financial and nonfinancial components were separated, and avoid counting the same change twice.
The difference between a balance rollforward and a performance explanation is important. An investor reading only the opening and closing balances could see a decline of 10 and infer that risks have disappeared. The illustrative detail shows that new business increased risk, while uncertainty releases and assumption changes more than offset the increase. The same net decline might also result from a model recalibration unrelated to improved claims experience. A meaningful disclosure separates these economic explanations.
The interaction with profit emergence
The CSM releases over coverage units as the entity provides insurance contract services. The risk adjustment is itself released as the entity is released from nonfinancial risk. Their release patterns need not coincide. A long-tail general liability contract may generate coverage service over an initial period and retain settlement uncertainty for years afterward. In contrast, some short-duration coverages may see exposure and settlement uncertainty fall together. As a result, the CSM and the risk adjustment can have different timing profiles even in the same accounting group.
Changes relating to future service can adjust the CSM rather than appear immediately in service profit, within the general model's limitations. The CSM cannot simply be reconstructed from the risk adjustment rollforward, because it also incorporates future-service changes in expected cash flows, accretion, currency effects, and the release based on coverage units. An analytically complete account therefore treats the CSM and risk adjustment as related yet distinct variables. The IFRS 17 general measurement requirements establish that general architecture.
The premium allocation approach
The premium allocation approach (PAA) simplifies the measurement of eligible short-duration contracts, particularly the liability for remaining coverage. The PAA's measurement simplifications should not lead readers to assume that nonfinancial-risk compensation disappears from the entire insurance accounting structure. The LIC under PAA generally continues to reflect fulfillment cash flows, including a risk adjustment for nonfinancial risk; the LRC can also require onerous-contract assessments. Paragraphs 53–59 of the Australian implementation of IFRS 17 make this distinction available in a clear primary-source text.
This has practical implications for comparisons. Two insurers specializing in one-year property policies may both apply PAA, yet their unpaid-claims liabilities can embody different nonfinancial risks because one writes fast-settling personal property claims and the other writes catastrophe-exposed commercial business. Disclosure analysis must therefore distinguish measurement approach from insurance line and from settlement profile. The existence of a simplified LRC calculation is not evidence of a negligible LIC risk adjustment.
Changes in discount rates and the disaggregation choice
The measurement of insurance contracts includes discounting and, where applicable, insurance finance income or expenses. Some nonfinancial-risk adjustments are determined using techniques that involve discounted amounts and finance effects. IFRS 17 requires disclosures of methods used to measure risk adjustment and whether changes are disaggregated between the insurance service and finance components. An entity's choice can affect the apparent volatility of the reported service result without changing the underlying claim distribution. The IFRS 17 paragraph 117 disclosures are crucial to interpreting performance across firms.
A reader should avoid equating a line called “risk adjustment release” with a pure decrease in insurance risk unless the entity explains how the finance effect is treated. Likewise, performance comparisons should consider differences in other comprehensive income choices, discount-rate presentation, and accounting methods relevant to insurance contracts. Economic risk may move similarly across two insurers while reported service and finance effects differ because the financial-statement presentation differs.
Quantitative Disclosure Requirements
The structure of the IFRS 17 disclosure obligation
IFRS 17's disclosure objective is to provide information that, together with the primary financial statements, enables users to assess the effect of insurance contracts on financial position, financial performance, and cash flows. Paragraphs 93–96 establish the general objectives, and subsequent provisions require reconciliation, significant judgment, uncertainty, and risk-exposure information. The EU-endorsed disclosure text integrates the risk adjustment into this broader reporting framework.
Paragraph 117(c)(ii) requires disclosure of the method used to determine the risk adjustment for nonfinancial risk, including whether changes in that adjustment are disaggregated between the insurance service component and the insurance finance component or presented in full in the insurance service result. Paragraph 119 specifically requires disclosure of the confidence level used in determining the adjustment. If an entity uses a different technique, such as cost of capital, it must disclose the technique and the equivalent confidence level. The paragraph 119 requirement thus imposes a common explanatory reference even where measurement techniques differ.
The equivalent confidence level is a translation of a reported margin into a percentile of the relevant distribution. It is not itself a mandatory choice of technique. For instance, an insurer may calculate an adjustment using a discounted capital-cost method and then determine which quantile-based loading would yield the same monetary amount for the modeled distribution. Another insurer may calculate the adjustment directly using a percentile. Their disclosed levels can be placed side by side, but a complete comparison must identify whether the loss distributions, risk exposures, aggregation levels, and assumed dependence are comparable.
Deriving an equivalent confidence level
Under the simple normal model previously introduced, the equivalence calculation is reproducible. If X has mean 100 and standard deviation 20 and a separately calculated adjustment equals 15, then the corresponding standard-normal score is 15/20 = 0.75. The standard normal cumulative distribution at 0.75 is approximately 0.7734, so the illustrative equivalent confidence level is about 77.34 percent. That statement is conditional on the distribution, horizon, and risk perimeter. It does not imply that a 15-unit adjustment universally represents 77.34 percent confidence.
Now consider an unchanged 15-unit adjustment but a revised estimated standard deviation of 30. The score becomes 0.5 and the equivalent confidence level becomes approximately 69.15 percent. The adjustment has not changed, yet the reported equivalent percentile would change because the modeled distribution changed. Conversely, a constant confidence level can produce a different monetary amount when the modeled uncertainty increases. Both relationships must be understood when interpreting financial-statement trends.
This illustrates a limitation of judging “prudence” from the reported percentile alone. A 70 percent equivalent confidence level for a highly skewed catastrophe portfolio and an 80 percent level for a stable mortality portfolio cannot be ranked mechanically by their numbers. The percentiles apply to different distributions and risk-bearing contexts. A useful disclosure also identifies the portfolios covered, model assumptions, material diversification, and whether the percentile is a single aggregate equivalent or varies by business line.
A cost-of-capital illustration
A cost-of-capital technique may relate the compensation for nonfinancial uncertainty to projected future required risk-bearing capital. In a deliberately simplified illustrative example, suppose projected relevant capital amounts for three annual periods are 100, 80, and 50; a constant annual capital cost is 6 percent; and illustrative discounting is 3 percent per year. Then:
The amount is not an IFRS-mandated risk adjustment, a Solvency II calculation, or a claim that 6 percent is an IASB-prescribed capital cost. It illustrates how an entity may form a risk-compensation estimate from projected risk-bearing capacity under explicit assumptions. Different timing conventions, capital attribution, diversification treatment, and cost assumptions can change the result. For this result to be translated into an equivalent confidence level, an additional probability distribution for the covered nonfinancial cash flows must be supplied.
For a methodological comparison, an exact replication should record the modeled cash-flow horizon, whether capital is pre- or post-diversification, the capital base, the discounting convention, and the translation procedure. The International Actuarial Note 100 guidance discusses recognized estimation approaches without imposing one approach on all entities.
Reporting granularity and aggregation
Paragraph 119 is often satisfied by an entity-level or business-segment-level equivalent confidence level, but the informational value of the disclosure depends on aggregation. A single figure may summarize a portfolio that combines catastrophe business, long-tail liability, term life, and annuity guarantees. Even if the total risk adjustment is appropriately measured, an aggregate percentile may conceal large variation among products. Conversely, extremely granular disclosures can bury material movements in detail. The useful reporting level is one at which the user can connect the confidence figure to the insured risks and relevant liability balances.
Three distinct aggregation questions should be kept visible. First, at what level are risks modeled and diversified? Second, at what level is the risk adjustment allocated for contract-group measurement? Third, at what level are the method and equivalent confidence level disclosed? Those levels may differ. A coherent financial statement makes the relationships understandable rather than relying on a single number to perform all three functions.
Where an entity reports only the aggregate adjustment, analysts can derive ratios such as RA/LIC, RA/LRC, or RA/discounted expected claims if the relevant denominator is available. Such ratios can be useful as descriptive statistics, but they are sensitive to business mix, contract duration, and measurement technique. They should not be interpreted as standardized risk premiums or direct measures of solvency. Their strength is diagnostic: a sharp unexplained change can identify a question worth asking.
Reconciliations, significant judgments, and risk sensitivity
A complete disclosure package should allow readers to link risk adjustment balances and changes with the broader insurance-liability reconciliations and the reported insurance service result. Significant judgments include estimation methods, risk appetite, included diversification, choice of confidence-level conversion, and disaggregation of finance effects. The standard also requires disclosure of information concerning risks, exposures, and sensitivities to insurance and market risk where applicable. Relevant risk-related disclosure provisions are located in paragraphs 121–132, including sensitivity analysis.
The sensitivity exercise should state what changes and what remains fixed. A reported increase in the adjustment might reflect an adverse loss-development forecast, an increase in volatility, a higher required confidence level, or a decrease in diversification. A sensitivity table isolating one variable at a time is analytically cleaner than presenting several simultaneous changes as if they were one cause. Where assumptions interact—for instance, more severe catastrophes also increasing correlations—the interaction should be described and, where material, quantified.
The interpretive minimum is therefore more demanding than a label and amount: the reporting entity should explain the technique, the equivalent percentile, the principal drivers of movement, important diversification treatment, accounting presentation, and the risk exposures to which the measure relates. This minimum is an analytical reading recommendation developed from existing requirements and observed limitations, not an assertion that IFRS 17 prescribes an identical tabular format for every insurer.
Financial Statement Sensitivity
Sensitivity to uncertainty and the reported margin
Holding the normal illustrative distribution's 75 percent confidence level fixed, raising the standard deviation from 20 to 30 raises the calculated adjustment from approximately 13.49 to approximately 20.23. The expected loss remains 100. This 6.74 increase results entirely from the assumed wider distribution. If the associated contract group remains profitable, the larger adjustment can reduce the initial CSM under the general measurement model. In later periods, revised uncertainty can influence subsequent service results and CSM according to the nature of the change.
Holding the standard deviation at 20 while increasing the confidence level from 75 to 85 percent raises the adjustment by approximately 7.24. These two sensitivities are different economically. The first changes the estimated uncertainty of the business; the second changes the required compensation represented by the measurement. The distinction helps analysts avoid treating every increase in risk adjustment as evidence of deteriorating underlying insurance performance.
The calculation also highlights disclosure comparability. A firm that increases the risk appetite charge may disclose an increasing adjustment even when the underwriting portfolio becomes less volatile. Another firm may disclose a stable adjustment while its modeled distribution widens, if the equivalent confidence level declines. An observed accounting time series without the percentile and model context is therefore incomplete evidence about risk evolution.
Dependence sensitivity
In the earlier two-portfolio illustration, the correlation of 0.25 produced a combined standard deviation of 40. If correlation rises to 0.75 with marginal standard deviations unchanged at 20 and 30, combined standard deviation becomes:
At 75 percent confidence the resulting aggregate loading is approximately 31.64, compared with approximately 26.98 under the lower correlation. The increase of about 4.66 is driven solely by greater dependence. This shows why disclosures about diversification and accumulation matter even when line-by-line frequency and severity estimates remain stable. The precise numerical impact would differ for skewed distributions and catastrophe tail dependence.
From a reporting perspective the effect could alter the risk adjustment without changing expected claims. A further allocation across groups may cause different group-level impacts. This dependence effect is especially relevant to large internationally diversified insurers where catastrophe, health, credit-related insurance, or litigation shocks may be correlated across ostensibly separate business segments.
Sensitivity to cash-flow timing
Two claim portfolios may have the same undiscounted expected loss but different liability durations. If the timing of expected payments becomes longer, the discounted mean can fall or rise according to rates and expected claim inflation. The separate nonfinancial risk compensation may also change because uncertainty persists longer, the range of settlement amounts widens, or a cost-of-capital calculation includes more future periods. These mechanisms must be traced separately. The IFRS 17 measurement guidance requires current, explicit, unbiased estimates of fulfillment cash flows and an appropriate treatment of uncertainty.
Suppose a 100-unit expected claim is now payable in three years rather than one. At an illustrative constant 3 percent discount rate, its expected present value changes from 97.09 to 91.51, a decline of 5.58. That mathematical decline says nothing by itself about the risk adjustment. The adjustment depends on uncertainty about amount and timing and the compensation required to bear it. A longer horizon could raise the adjustment even while the discounted mean declines. Financial-statement users need both parts of the calculation to understand the total obligation.
The risk adjustment and accounting performance
The risk adjustment's effect on profit is conditional on contract profitability and the allocation of changes to service periods. An increase at initial recognition can reduce the CSM on a profitable group. An increase related to past service can affect current service performance, while a future-service change may adjust the CSM where the general model permits. The reporting effect is consequently a function of the same economic change's accounting location. The measurement and service result requirements establish those distinctions.
If two insurers use different methods of distinguishing service and finance components of the adjustment, their service-result volatility can differ even if the total fulfillment-liability movement is broadly similar. This raises an analytical caution for performance comparisons: a lower service-result volatility is not itself evidence of lower underlying risk. The notes must be read alongside accounting-policy disclosures and liability reconciliations.
Separate presentation under financial and prudential frameworks
Prudential Solvency II valuation has a statutory risk margin calculated using projected capital needs; IFRS 17 reports entity-specific compensation for nonfinancial uncertainty. The levels cannot be converted into one another by applying a single fixed multiplier. EIOPA's 2024 insurer implementation evidence found significant differences between IFRS 17 adjustments and Solvency II margins and showed variation by business type and measurement approach. This finding supports regime-specific reconciliation, not claims that one framework systematically measures insurance risk more correctly.
The legal chronology is also relevant. At the November 2026 research cutoff, Solvency II reforms had been adopted, including changes to the risk-margin cost-of-capital rate and a time-dependent capital adjustment. The 2025 amendment to Solvency II and Delegated Regulation 2026/269 provide for a 4.75 percent rate from January 30, 2027, rather than treating that rate as already applicable in November 2026. EIOPA's July 2026 revised valuation guidance likewise states the January 30, 2027 applicability date. Any comparison using an earlier 6 percent regulatory cost rate must identify the historical reporting period. These prospective changes are a reason to time-stamp comparative studies carefully, and do not modify IFRS 17's own definition of the adjustment.
IFRS 18 and upcoming presentation changes
IFRS 18, which replaces IAS 1 for relevant reporting periods beginning on or after January 1, 2027, affects the general architecture of financial-statement presentation and disclosure. The IFRS 18 issued-standard overview states the effective date and permits earlier application under the standard's requirements. In an insurer's reporting architecture, the transition can affect how performance information is organized, although IFRS 17 continues to govern insurance contract measurement.
As of November 2026, an analysis of 2025 annual reports must evaluate reporting under the standards applicable to those reports; it must not retrospectively assume general mandatory application of IFRS 18. Similarly, a forward-looking discussion of January 2027 changes must be labeled as prospective. These date distinctions are substantive requirements for scholarly comparison.
International Implementation Evidence
The empirical questions
Three empirical questions guide comparison of risk adjustment practices. Which estimation methods are used, and at what levels? How consistently are equivalent confidence levels and disaggregation methods disclosed? What can be learned from reported ratios or comparisons with prudential margins when business mix and regulatory differences remain? Evidence should be evaluated according to sampling frame, year, region, type of insurer, and data availability. A study based on large listed European groups cannot be presumed representative of all mutual insurers, emerging-market insurers, or smaller national insurance markets.
The international implementation period also matters. IFRS 17 became effective for many IFRS reporters in 2023, and early studies examined transitional disclosures. Subsequent reporting in 2024 and 2025 incorporated refinements in accounting policies, tables, and analyst communication. A difference observed between studies may reflect changes in practices, changes in samples, or changes in the reporting period; the studies are not automatically successive measurements of one fixed population.
EIOPA's first implementation study
EIOPA's April 2024 IFRS 17 implementation report examined a sample of 53 European insurance and reinsurance groups from 17 European Union Member States, drawing on early IFRS 17 reporting. Its findings showed diversity in the techniques used to determine risk adjustment and in the relationship between accounting adjustments and Solvency II prudential risk margins. The reported methods included value-at-risk techniques among 60 percent and cost-of-capital techniques among 47 percent of respondents. These percentages can exceed 100 percent in sum because some respondents used more than one technique; interpreting them as exclusive market shares would be incorrect.
The study reported equivalent confidence levels generally in a range around 75 to 85 percent and averaging around 80 percent for life and nonlife business within the surveyed population. It also identified differences in the treatment and measurement of diversification and noted that some entities achieved methodological synergies with existing Solvency II work. These observations provide specific support for the argument that the required disclosed percentile is helpful but does not make methodologies or risk perimeters uniform.
EIOPA's analysis compares adjustments and margins across different business types. Its approximate sample ratios show that the IFRS 17 adjustment can be higher than the Solvency II risk margin in one segment yet lower in another. The study's 2024 data and then-existing Solvency II parameters should be preserved when reporting those comparisons. Later prudential reforms effective in 2027 change a relevant comparator and require a separate prospective analysis.
Academic examination of first-year European reports
Ter Hoeven, Borelli, and van Vlijmen (2024) examined the first year of IFRS 17 application in the financial statements of 24 major European insurers. Their peer-reviewed comparative study documents implementation and disclosure issues alongside the structural changes introduced by the standard. Its sample design offers useful evidence about the financial-reporting behavior of large public insurers, but it does not establish how all entities worldwide used confidence-level techniques or risk-adjustment assumptions.
The significance of first-year academic work lies partly in its attention to accounting comparability. The new standard can standardize the basic recognition and measurement vocabulary while allowing substantial choices in implementation. Analysts therefore need to distinguish formal consistency in headings and statement lines from substantive consistency in assumptions and technique. This methodological distinction also applies to statistical comparisons of reported risk adjustment percentages.
The study's year-one observations should not be transplanted automatically to the 2025 reporting population. Differences in sample composition and reporting maturity may explain apparent changes. A stronger longitudinal design would follow the same insurers across several years, distinguish restatements from genuine model revisions, and assess the stability of equivalent confidence levels alongside risk-adjustment movements.
Third-year global observation
KPMG's April 2026 survey of 55 insurers' 2025 annual reports examines accounting policies, significant judgments, and disclosure practices under IFRS 17 and IFRS 9. It identifies increasingly mature reporting practices while emphasizing continuing comparability challenges. Its detailed risk-adjustment analysis documents variation in the reporting of equivalent confidence levels and in the disaggregation of risk adjustment between service and finance components.
This international survey is valuable for its later reporting date and broader regional reach, but its insurer selection is not a random world census. Differences in business mix, reporting basis, reporting date, and national adoption can change the observed distribution of disclosures. The evidence supports a claim about persistent heterogeneity among the entities examined, rather than a precise claim about the percentage of all global insurers following any one disclosure practice.
The same study's presentation of confidence-level disclosures by insurer segment illustrates a key analytical issue: a single consolidated equivalent level can be less comparable than a well-explained line-of-business level, because different segment weights can change the aggregate even when underlying business-specific confidence assumptions do not change. A reporting improvement would connect the confidence-level calculation to the risk-adjustment balance and the model's diversification perimeter.
Actuarial practitioner observations
The Actuaries Institute's 2025 post-implementation survey, published in 2026, examines IFRS 17 implementation experiences in Australia and New Zealand. The survey's responses indicate the coexistence of cost-of-capital and confidence-level approaches and provide useful practitioner evidence about implementation choices. Survey results reveal professional practice among respondents; they do not determine IFRS 17 compliance, and they should not be generalized beyond their sampling frame.
The study is especially useful as a counterpoint to European evidence. Institutional practices and product structures differ across jurisdictions, even when IFRS 17 terminology is shared. Across surveys, a method's prevalence cannot be compared without specifying whether respondents were permitted to report multiple methods, whether the unit is a company or a segment, and whether “cost of capital” refers to identical estimation conventions. These details decide the meaning of reported percentages.
Comparative interpretation and observational limits
The combined evidence supports several substantive observations. First, the legal and accounting definition of risk adjustment has become common among adopters of IFRS 17, while its technique remains diverse. Second, regulators' and analysts' access to equivalent confidence levels has improved the availability of a comparable descriptor, but meaningful comparison remains conditional on risk perimeter, diversification, model calibration, and reporting aggregation. Third, the observed connection between IFRS 17 and Solvency II work is often operational and computational, without making the two accounting and prudential outputs equivalent. Fourth, disclosures become more informative when methods, period movements, and risk exposures are explained together.
These are qualitative syntheses from a standards text and selected observational studies. They are not estimates of causal effects, and the available studies are not a unified representative cross-country panel. There is no basis in these sources to assert that a single confidence level is optimal worldwide or that a particular risk-adjustment method produces superior financial performance. An insurer's chosen method should be evaluated in relation to the compensation the entity requires, the model's evidence, and the transparency of the resulting disclosures.
Implications for research and reporting practice
Future empirical work can strengthen comparison by assembling entity-level time series that consistently separate gross issued-contract adjustments, reinsurance-held adjustments, and measurement approaches; by matching disclosed equivalent confidence levels to changes in claim development and model assumptions; and by analyzing how diversification practice relates to capital, portfolio mix, and jurisdiction. Such research would need data controls for currency, inflation, discount rates, consolidation boundary, and transition accounting. It would also require disclosure-level measurement of missing information rather than treating a nondisclosed metric as zero.
For readers of current reports, a structured sequence is useful: identify the insurance liabilities being measured; read the risk-adjustment definition and method; interpret the equivalent percentile in the context of the modeled portfolio; reconcile the opening and closing adjustment; trace future-service, current-service, and finance effects; examine risk sensitivities and diversification; then compare with peer disclosures of similar contract portfolios. This reading procedure is an analytical synthesis of established concepts and observed disclosure problems, not a new doctrinal standard.
Conclusion
The risk adjustment for nonfinancial risk is an explicit measurement of the compensation an insurer requires for bearing uncertainty about insurance contract cash flows. In IFRS 17 it belongs to the fulfillment cash-flow architecture and interacts with the contractual service margin, the liabilities for remaining coverage and incurred claims, reinsurance held, and the reported insurance service and finance results. Its accounting meaning is more precise than any single computational technique, because the standard defines the compensation being measured while permitting techniques consistent with that definition.
The central disclosure requirement is correspondingly twofold. Users need the amount that affects the statement of financial position and income; they also need sufficient explanation to understand how that amount relates to nonfinancial uncertainty. Disclosure of the measurement method and its equivalent confidence level under paragraph 119 is necessary for comparability. The full interpretive picture also includes risk perimeter, diversification, model assumptions, measurement approach, accounting movements, and sensitivities. The arithmetic examples establish why a single percentage cannot communicate these facts by itself: identical margins can represent different equivalent percentiles under different distributions, and identical percentiles can imply different monetary adjustments.
International implementation evidence confirms that methodological and disclosure choices vary. EIOPA's 2024 European observations, peer-reviewed examinations of first-year reporting, KPMG's 2026 study of 2025 financial statements, and professional surveys show how a common standard coexists with heterogeneity in calculation and explanation. Their samples are informative but do not justify universal claims about all insurers. The contrast with the Solvency II risk margin further confirms the importance of comparing like with like and accounting for regulatory effective dates, including the January 2027 changes adopted before this research cutoff.
The substantive conclusion is that the analytical value of risk adjustment lies in the relationship between measured compensation, underlying distribution, recognized diversification, and recognized period movement. A well-supported disclosure turns a numerical balance into an interpretable statement about the insurer's remaining exposure to nonfinancial uncertainty. The monograph establishes this conclusion through existing primary standards, published empirical evidence, and transparent calculations. It introduces no new named theory or accounting doctrine.
References
Actuaries Institute. (2026). IFRS 17 post-implementation 2025 survey report. Sydney: Actuaries Institute. https://content.actuaries.asn.au/resources/resource-ce6yyqn64sx3-2093352434-60769
Australian Accounting Standards Board. (2026 compilation). AASB 17 Insurance Contracts (compilation applicable July 2026 to periods before January 2027, as identified on the official standards site). https://standards.aasb.gov.au/aasb-17-dec-2022-0
European Commission. (2021). Commission Regulation (EU) 2021/2036 of 19 November 2021 amending Regulation (EC) No 1126/2008 as regards International Financial Reporting Standard 17. Official Journal of the European Union. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32021R2036
European Commission. (2026). Commission Delegated Regulation (EU) 2026/269 of 29 October 2025 amending Delegated Regulation (EU) 2015/35 as regards technical provisions and other Solvency II matters. Official Journal of the European Union, February 18, 2026. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AL_202600269
European Insurance and Occupational Pensions Authority. (2024). IFRS 17—Insurance contracts report (April 15, 2024). Frankfurt am Main: EIOPA. https://www.eiopa.europa.eu/publications/report-implementation-ifrs-17-insurance-contracts_en
European Insurance and Occupational Pensions Authority. (2026). Final report on revised guidelines on valuation of technical provisions—Solvency II Review (July 15, 2026). https://www.eiopa.europa.eu/publications/final-report-revised-guidelines-valuation-technical-provisions-solvency-ii-review_en
European Insurance and Occupational Pensions Authority. (2026, accessed November 14). Solvency II Single Rulebook: Delegated Regulation (EU) 2015/35, Article 37, Calculation of the risk margin. https://www.eiopa.europa.eu/rulebook/solvency-ii-single-rulebook/article-2450_en
European Parliament and Council of the European Union. (2025). Directive (EU) 2025/2 of 27 November 2024 amending Directive 2009/138/EC and related directives. Official Journal of the European Union, January 8, 2025. https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX%3A32025L0002
IFRS Foundation. (2026, accessed November 14). IFRS 17 Insurance Contracts: Issued standard overview. London: IFRS Foundation. https://www.ifrs.org/issued-standards/list-of-standards/ifrs-17-insurance-contracts/
IFRS Foundation. (2026, accessed November 14). IFRS 17 Insurance Contracts: Key terms. London: IFRS Foundation. https://www.ifrs.org/supporting-implementation/supporting-materials-by-ifrs-standards/ifrs-17/key-terms/
IFRS Foundation. (2026, accessed November 14). IFRS 18 Presentation and Disclosure in Financial Statements: Issued standard overview. London: IFRS Foundation. https://www.ifrs.org/issued-standards/list-of-standards/ifrs-18-presentation-and-disclosure-in-financial-statements/?hl=en-GB
International Actuarial Association. (2021). International Actuarial Note 100: Application of IFRS 17 Insurance Contracts (August 31, 2021). Ottawa: IAA. https://actuaries.org/app/uploads/2025/04/IAA_IAN100_31August2021.pdf
KPMG International. (2026). Insurers' 2025 annual financial statements: Updated observations from insurers' reporting under IFRS 17 and IFRS 9 (April 21, 2026). https://kpmg.com/xx/en/our-insights/ifrg/2026/2025-annual-reports-real-time-ifrs17.html
KPMG International. (2026). Real-time IFRS 17: Insurers' 2025 annual reporting under IFRS 17 and IFRS 9. KPMG IFRG technical publication. https://assets.kpmg.com/content/dam/kpmgsites/xx/pdf/ifrg/2026/isg-talkbook-real-time-ifrs-17-2025-annual-reporting.pdf.coredownload.pdf
Official Journal of the European Union. (2021). Commission Regulation (EU) 2021/2036, English adopted IFRS 17 text and application guidance. https://eur-lex.europa.eu/eli/reg/2021/2036/oj/eng
ter Hoeven, R., Borelli, A., & van Vlijmen, P. (2024). First year's application of IFRS 17 in the financial statements of European insurance companies. Maandblad voor Accountancy en Bedrijfseconomie, 98(6), 403–422. https://doi.org/10.5117/mab.98.139090