Optimising AASB 17 PAA Eligibility Testing

With the conclusion of the first year of reporting on an AASB 17 basis, the focus for insurers has shifted from implementation to refinement and optimisation.

The significant volume of information provided by public disclosures of audited financial statements allows for useful comparison and benchmarking, and an ASTIN Sub-Committee of the International Actuarial Association has been formed to investigate best practices on IFRS 17 at a global level.

At a local Asia-Pacific level, auditing firms have published comparative studies analysing the implementation of AASB 17/IFRS 17 (for example IFRS  17:First look at FY23 disclosures in Australia and A comparative analysis of initial disclosures ).

One key theme that has emerged is that, despite IFRS 17's aim of standardising the accounting disclosure requirements across insurers, there has been a multitude of approaches taken by insurers in the underlying modelling of the financials which vary by granularity, sophistication and calibration.

As insurers analyse the disclosures of their peers, opportunities for enhancements naturally arise as insurers refine and align their approaches to leading industry practices. One area where we have seen different adopted modelling methodologies relates to testing the Premium Allocation Approach (PAA) eligibility.

With the accounting standard being principle-based, we have seen a variety of approaches that vary in both the underlying modelling methodology and the adopted assumptions and accounting policies.

In many cases, an insurer has opted to apply paragraph 53(a) in PAA eligibility testing and show that the PAA simplification would produce a Liability for Remaining Coverage (LFRC) for the group of contracts that would not differ materially from the application of the General Measurement Model (GMM).

However, the insurer's modelling approach has not always made it a simple task to identify and quantify differences that arise between the GMM and PAA.

Whilst this may have been a relatively minor issue during initial implementation where significant time was spent in determining PAA eligibility, materiality policy and all other IFRS 17 implementation related activities, a more transparent approach will assist in streamlining operations as AASB 17 implementation is absorbed into business-as-usual activities.

To this end, the British Actuarial Journal has recently published our paper on optimising the PAA eligibility testing process (determining the mathematical conditions under which the PAA and GMM produce identical LFRC), which represents the culmination of the many refinements we have identified when conducting PAA eligibility testing both in Australia and overseas.

A mathematical framework for analysing PAA eligibility

Our paper details a systematic mathematical framework for identifying and quantifying differences that arise between the GMM and PAA. 

This is achieved by firstly laying out the modelling assumptions and methodology which would give rise to the same LFRC balance determined under the GMM and PAA (which we call the sufficient conditions model).

Overall, we have identified over a dozen sufficient conditions, covering stipulations on the insurance contracts, discount rate, coverage units, cashflows, risk adjustment, revenue recognition and the pattern of amortising the insurance acquisition cash flow.

Whilst the underlying idea may not be complex (i.e., to match the balance and revenue/expense recognition produced by the GMM with those from the PAA), we have found the details to be non-trivial, so we have provided in the paper explicit formulae for each assumption, together with worked examples and discussion of the difficulties faced by practitioners.

Next, by comparing each assumption adopted by the insurer with the assumptions from the sufficient conditions model, the drivers behind the discrepancies between the GMM and PAA can be identified and quantified. The value provided by pinpointing the sources of discrepancies allows for a more tractable and transparent understanding of the risks so that attention can be directed at resolving the material issues.

Apart from increasing the transparency in identifying and understanding sources of discrepancies in the GMM and PAA, the mathematical framework also offers AASB 17 practitioners the following set of opportunities:

• The framework provides a mathematical lens on which to refine accounting policies, where the objective is to minimise the gap between the GMM and PAA when computing the LFRC. In particular, the quantification provided by this exercise may also lead to refinements in the insurer's materiality policy.
• The framework has great potential in reducing the amount of tedious recalibration when performing PAA eligibility testing. This is because a large part of the testing could be reduced to observing whether the sufficient conditions have been satisfied (a much simpler task) compared to a less scalable approach of a full recalibration every time an assumption is tweaked.
• Provided the group of insurance contracts under consideration are reasonably homogenous and the economic environment reasonably stable, our paper suggests a number of practical expedients (in relation to contracts added to the group after initial recognition and calibrating discount rates), which could offer insights into techniques for simplifying a complex PAA or GMM model.
• Whilst the framework does not (nor aims to) provide an automatic approach to pass PAA eligibility testing under all circumstances (indeed paragraph 54 requires an insurer to test for significant variability in cash flows), our experience (also supported by anecdotal evidence from our colleagues) has suggested the framework provides a dependable and systematic basis for testing for paragraph 54. In general, we have found the overall discrepancies to be smaller when using the framework to test for paragraph 54 compared to alternative approaches.
• Whilst the framework focuses on direct insurance contracts, the ideas (with suitable modifications and subject to complexity of contracts) can be extended to reinsurance contracts, thereby allowing for PAA disclosures to be presented, which are less onerous than the GMM counterpart. We have observed cases where contracts issued for adverse development cover have been measured using PAA, after a number of optimisation strategies from our work were implemented.

The transition from initial implementation to an optimisation phase presents a unique opportunity for insurers to enhance operational efficiency and implement best practices, which not only satisfies regulatory requirements, but also reduces the complexity and resource intensity in its implementation.

Our work on optimising the PAA eligibility testing process is one such example of a systematic and transparent framework offering both enhanced insights for management and strategies for streamlining operations.

Indeed, we have observed enhancements made in a much broader accounting modernisation context, including harmonising data-actuarial-finance architecture, technology and processes for automated and faster insights, and investment into both financial and management reporting. As IFRS 17 enters a more business-as-usual context, optimisation will offer insurers a source of competitive advantage.

Lee T, Jagga A. (2024). Sufficient mathematical conditions for identical estimation of the liability for remaining coverage under the general measurement model and premium allocation approach, British Actuarial Journal, available at https://doi.org/10.1017/S1357321724000242 (accessed 9 January 2025).

PwC Australia (2024). IFRS 17: First look at FY23 disclosures in Australia, available at https://www.pwc.com.au/insurance/ifrs-17.html (accessed 9 January 2025).

PwC Singapore (2024). IFRS 17 implementation in Singapore: a comparative analysis of initial disclosures, available at https://www.pwc.com/sg/en/publications/ifrs-17-implementation-in-singapore-a-comparative-analysis-of-initial-disclosures.html (accessed 9 January 2025).