Making the transition to IFRS 17

Published in: IFRS 17

Companies: Milliman

Insurers must soon make some crucial decisions in how they will make the transition from current insurance contract accounting rules to IFRS 17. Henny Verheugen, principal and consulting actuary at Milliman, discusses the possible implications of each of the three approaches to transition, and why insurers will need to choose their options carefully

The publication in May 2017 of the long-awaited standard on insurance contracts, IFRS 17, begins the countdown for implementation. Begining on or after January 2021, entities will need to apply IFRS 17 for external reporting. Although early application is permitted in some circumstances, this timetable means that the first IFRS 17 compliant results will be published in Q1 2021 with comparatives needed for the previous period. Consequently, 1 January 2020 is the transition date. Transition and effective dates will evolve as per figure 1.

The decisions made at transition will define the future presentation of results for the existing and future new business of all entities, but especially for those primarily managing run-off business. It is important for entities to analyse in detail the consequences of their various accounting decisions on their future results.

Figure 1: IFRS17 transition and effective dates

Figure 1: IFRS17 transition and effective dates

Relation with IAS 8

In accordance with IAS 8, implementation of a new accounting policy must be done retrospectively as if IFRS 17 had always applied. The most challenging changes are for accrual type balances like the Contractual Service Margin (CSM) and accumulated Other Comprehensive Income (OCI), which are developed based on historical information.

The CSM represents the expected profit to be realised as the insurer provides service and is recognised over the term of the portfolio. The CSM also serves as a buffer for changes to estimated cash flows related to future services and thus reflects the impact of historical changes at that moment in the future fulfillment cash flows.

Insurance liabilities are to be measured using current assumptions at each valuaton date including discount rates. Entities are allowed to report the effect of changes in discount rates either in profit and loss or in OCI.

The impact of the retrospective application of IFRS 17 is reflected in the equity account of the entity.

Entities that only manage historical claims don't have to determine the CSM, because claims reserves don't contain a CSM. For those companies, the main focus will be on the calculation of the amount of OCI.

Accounting decisions

Two important decisions insurance entities have are to define the level of the risk adjustment (RA) and the circumstances for recognising the amounts in OCI. The level of RA needs to reflect the price an entity would charge to turn an uncertain cash flow into a certain one, or as stated in IFRS 17 the compensation that the entity requires for the uncertainty about the amount and timing of the cash flows.

However, the requirement is not described in great detail and leaves some freedom for the entity. Decisions about the level of the RA, its impact on the CSM and the use of the OCI need to be made before the transition. Each decision will have specific consequences.

For example, entities will see an immediate impact on the level of the CSM at the transition date based on their approach to determine the level of RA. A high RA will lead to a low or even nil CSM, thereby reducing the capacity of the CSM to absorb changes in the value of future fulfilment cash flows. On the other hand, a high RA may give shareholders a stable stream of earnings because the RA will follow a release pattern that is consistent with how the entity is released from the non-financial risk.

The decision to use OCI for reporting the impact of discount rate changes is often used in conjunction with the classification of financial instruments at Fair Value through OCI (FVOCI).

The FVOCI option is only applicable for financial instruments that meet the Solely Payments of Principal and Interest test specified in IFRS 9. By selecting the FVOCI option, it will be possible to match the movement in the OCI for the change in discount rates applicable to the insurance liabilities.

It is important to note that the amount of OCI on the balance sheet for financial instruments can flow into the income statement if the financial instruments are sold. For insurance liabilities, it will be more difficult to realise the OCI. For insurance liabilities, OCI will be realised over time as the liabilities run off, but the entity doesn't have many means to change the pattern.

Three approaches

IFRS 17 contains a complete appendix (C) about the transition in which three approaches are presented. A high level overview is depicted in Figure 2.

The approaches are described in more detail in the following paragraphs.

Figure 2: The three approaches to IFRS17 transition

Figure 2: The three approaches to IFRS17 transition

Full retrospective approach

The required presentation for certain items on the balance sheet at transition date (1 January 2020) are summarised in Figure 3.

The determination of the BE and RA is based on forward-looking information as per the transition date. This is the same approach (including models, assumptions, etc) that is used after the transition as part of the ongoing reporting process. The discount rate effect and the contractual service margin require a complicated retrospective analysis that can be burdensome to implement.

The discount rate at initial recognition date and all assumed changes between the initial recognition date and transition date will be required to determine the level of the CSM for a group of policies.

An additional complication of the full retrospective approach is that the annual historical cohorts need to be used, which is only possible if the complete history including all the modifications of the group of contracts is available. All historical assumptions like mortality, morbidity and lapse rates as well as expense assumptions need to be available.

The next section summarises alternative approaches that are allowed when an entity can demonstrate that the full retrospective approach is impracticable.

Figure 3: Required presentation on the transition date balance sheet following the full retrospective approach

ItemsFull retrospective approach
Best estimate of expected cash flows (BE) Expected present value using the discount curve and BE assumptions as per the transition date.
Risk adjustment (RA) Based on assumptions and entity's view as per the transition date.
Contractual service margin (CSM) Based on the calculations of the BE and RA at the initial recognition date and consideration of account developments to the transition date.
Discount rate effect Difference in expected present value of the insurance portfolio using the yield curve and BE assumptions at the transition date minus the expected present value using the curve on the initial recognition date and the BE assumptions as per the transition date. Note that this effect can be reflected in the CSM (if the variable fee approach applies), in the OCI or in the statement of profit and loss.

Practicability of the full retrospective approach

According to IAS 8, a full retrospective approach is required to determine the financial position for the earliest prior period presented, but the standard also describes certain circumstances for determining what constitutes impractability as:

  • The effects of the retrospective application are not determinable
  • The retrospective application requires assumptions about what management's intent would have been in that period
  • The retrospective application requires significant estimates of the amounts, but it is impossible to objectively distinguish information about those estimates that:
    • Provides evidence of circumstances that existed on the date(s) at which those amounts are to be recognised, measured or disclosed
    • Would have been available when the financial statements for that prior period were authorised for issue from other information

Hindsight should not be used when applying the new IFRS 17 standard to a prior period. However, IFRS 17 does not require entities to undertake exhaustive efforts to obtain objective information for the full retrospective approach. Entities should take into account all objective information that is reasonably available.

It is our expectation that for contracts that have an inception date in 2017 or later, it will be difficult to evidence that the full retrospective approach is impracticable. Starting 2017 entities are familiar with IFRS 17 and its requirements and hence can anticipate on collecting the information for the full retrospective approach.

When an entity cannot practicably determine the cumulative effect of applying the new IFRS 17 to all prior periods, it is allowed to apply a simplified approach.

It is permissible to choose at a group of contracts between a modified retrospective approach and a fair value approach to apply the new IFRS 17 standard. The choice of approach depends on which method better reflects the entity's results. Several relevant considerations are discussed in the following sections.

Modified retrospective approach

Under the modified retrospective approach, the goal is to achieve the closest outcome to the full retrospective application that is possible using reasonable and supportable information at the transition date. The entity has to maximise the use of information that would be used in a full retrospective application, where such information is available without undue cost or effort.

Instead of reconstructing all the historical changes in the CSM, the modified retrospective approach can set expected cash flows equal to the realised cash flows between the initial recognition and the transition date. Effectively, all historical cash flows were estimated perfectly until the transition date. A more complete summary of the modifications that are permitted are shown in Figure 4.

In making the determination to use the modified retrospective approach, entities should also consider the following practical constraints:

  • Grouping of contracts may be the only feasible solution because financial administration systems/general ledger information don't register the initial recognition date of the group for each cash flow. Also, allocation of expenses to the different years of initial recognition is likely impossible.
  • When the option of grouping of contracts from multiple years is applied, the CSM will be calculated at initial recognition of the earliest contract in the group. This may lead to significant timing differences between the real inception dates of insurance contracts and the recognition date of the group of contracts.
  • The yield curve may be determined based on the initial recognition date. Many in-force long-term contracts were issued when interest rates ranged between 4% and 8%. The yield curve at recognition for these contracts was materially higher than the current yield curve. The application of the high yield curve will lead to a relatively high CSM at recognition. Entities need to be aware that using a high yield curve will lead to higher future insurance finance expenses.
  • In order to determine the CSM at initial recognition, it will be necessary to calculate the time value of options and guarantees per that date. For many types of profit sharing contracts, stochastic calculations need to be made using economic scenarios calibrated on the historical economic parameters, which is an extremely challenging task.

The modified retrospective approach is not a simple solution given the level of evidence required to establish its correctness and the difficulties in collecting the appropriate historical information.

Figure 4: permitted modifications

ItemFull retrospective approachModified retrospective approach
Expected cash flows at the date of initial recognition Based on the calculations of the BE and RA at initial recognition date. Combination of
1) Realised cash flows that are known to have occurred between the date of initial recognition and the transition date
2) Expected cash flows by using current assumptions as per the transition date
Discount rate at the date of initial recognition Based on the market conditions at initial recognition date. Use actual yield curve for last three years and prior period yield curves based on calibration for last three years to observable index The appropriate average spread should be determined over the three years immediately preceding the transition date.
Alternatively, the yield curve at transition may be applied.
Risk adjustment at initial recognition Based on the assumptions and view of risk at initial recognition date. Estimate the RA at initial recognition by adjusting the RA determined at the transition date with the expected release between the initial recognition date and the transition date.
Grouping of contracts Contracts issued no more than a year apart which have similar risk characteristics and are managed together can be grouped together with at least three sub-groups based on expected profitability.

Contracts issued more than a year apart can be grouped together.
Grouping of contracts and assessing the applicability of the variable fee approach can be done at contract inception date or at the transition date.

Developments between the initial recognition date and the transition date impacting the CSM Reflect all developments between the initial recognition date and the transition date. Estimate CSM at the transition date reflecting the cash flows known to have occurred between the date of initial recognition and the transition date.
In case of grouping of contracts issued more than one year apart, there is an option to apply the discount rate at the transition date instead of the discount rate per inception to determine the CSM accrual and adjustments.

Fair value approach

The other alternative is the fair value approach. Under this methodology, the contractual service margin at the transition date is determined as the difference between the fair value (computed in accordance with IFRS 13) of the insurance contract and the present value of fulfilment cash flows (determined in accordance with IFRS 17). The fair value may differ from the fulfilment cash flows for a number of reasons including:

  • The fulfilment cash flows do not reflect the non-performance risk of the entity that issues the contract, whereas it is included in the fair value.
  • The fulfilment cash flows exclude overhead expenses which are not directly attributable to the contracts, whereas an overhead allowance is included in the fair value.
  • The discount curve used for the fulfilment cash flows should exclude the effect of any factors that influence observable market prices but that are not relevant to the cash flows of the insurance contract.
  • The BE cash flows and the RA reflect the entity's perception of risks, taking into account the entity's diversification benefit and degree of risk aversion, whereas the fair value is based on the exit value principle (the view of a typical market participant).
  • The contract boundary applied in an exit value situation may be different from the the contract boundary definition under IFRS 17. Generally M&A transactions take renewals and extensions of contracts into account. Although, the objectives of the fair value in an M&A and for the transition may be different, it may lead to differences in the future cash flows between the methods and have an impact on the CSM.
  • In the case of an exit value calculation, the purchasing entity will require a profit margin and hence, will decrease the fair value.
  • The fair value leads to unintended results as profitable business will lead to a relatively low fair value and low or zero CSM, while less profitable or onerous contracts will lead to a relatively high fair value and CSM. Given that entities may have more options for the transition, this unintended effect can be used to optimise the CSM.
  • IFRS 13 requires that the fair value of a financial liability with a demand feature is not less than the discounted amount payable on demand, whereas such a floor does not exist for fulfilment cash flows. This requirement is not applicable when the fair value is determined for the transition.

The grouping requirements for the fair value approach are the same as the modified retrospective approach.

The application of the fair value approach for the transition has several debatable topics, but it should be easiest to apply. The determination of the CSM under the fair value approach is a good starting point for the transition project and provides entities with a benchmark. For groups of contracts where the entity has the impression that the CSM is underestimated, entities can explore the modified retrospective approach.


Implementing IFRS 17 will touch on every corner of an entity's financial operation from systems and processes to controls and review. Each approach carries with it vastly different implications – some apparent, some perhaps unintended – for entities that will need to make fundamental decisions regarding their reporting, which ultimately will affect the way in which analysts perceive the entity.

Deciding on a way forward will require careful consideration of not only the financial position at transition but how earnings will emerge in the future from both in-force and new business.