IFRS17 actuarial models

Should you upgrade your actuarial modelling software for IFRS17?

For most of the actuaries out there the fun part is most definitely the actuarial models. Being from the actuarial field, this is a particular favourite of mine as well. IFRS17 gave insurers an interesting choice, do you improve on your current models or do you upgrade your system and buy something off the shelf that can more easily handle the new modelling requirements. Sometimes the best approach is probably to buy a new solution and other times to adjust what you already have (based on how mature and flexible your current models are). But where is the line where one has to decide to actually buy new actuarial software? Let’s discuss this by looking into more detail, the impact that IFRS17 has on your actuarial models and the types of questions we’ll need to ask ourselves.

Analysis of change and granularity of your actuarial models

The first component of the actuarial models that’s worth discussing is the analysis of change, or run order, that is currently used to generate expected cash flows. IFRS17 requires you to have, at a minimum, changes that unlock your CSM and those that don’t. On top of that, the changes that unlock your CSM should be calculated based on locked-in discount rates (for products falling under GMM). This alone is quite a large change that’s required, but not too hard to implement if you already have advanced modelling software (and access to the specialised resources that could do the work). Another complication is that cash flows may need to be calculated in more detail, for example, benefit outgo needs to be split out into those funded by the underlying item and funded by shareholders money.

Can my current models cater for new run orders using different financial assumptions?

Run order of your IFRS17 accounting engine

Some insurers don’t just want to go the minimum compliance route and actually do want to split out the impacts on projected cash flows into more detail. For these insurers, it could be a more complex process to generate inputs for your IFRS17 accounting engine (like your projected cash flows and coverage units) as now you need to ensure your analysis of change (and impacts you’d like to show separately) can actually be catered for in the IFRS17 accounting engine that you have chosen (or still need to choose). Most vendors are trying to implement an analysis of movement that can cater for all insurers, however some have not fully built in this functionality.

Does my IFRS17 accounting engine cater for the run order I’d like to implement in my actuarial models?

Level of aggregation

Now that insurers need to calculate a CSM, at IFRS17 group level, there’s a need for more granular cash flow projections and flexibility on aggregation of cash flows. An insurer can decide whether to output per policy results, which is aggregated to IFRS group level outside the actuarial software, or whether to output results at IFRS17 group level. The former provides more flexibility, but significantly increases runtime and hard disk storing space. The latter is possible if your profitability testing is well defined and will make it much simpler to include future profit projections in your calculations.

What’s the lowest level at which my actuarial models can currently produce projected cash flows and can they aggregate cash flows easily over different levels of granularity?

Profitability testing

As noted in a previous article, "International IFRS17 implementation status update", where an insurer does not have reasonable and supportable information to group contracts into sets of contracts, profitability testing may need to be performed on an individual contract-by-contract basis. This would be an exercise which will have to be automated, to tag policies to the groups to which they belong, and needs to be integrated with the actuarial models. Thus, when designing the actuarial models you’ll need to consider how you’ll use your models to inform profitability testing (where required). 

Do my actuarial models have the ability to cater for grouping of contracts and performing profitability testing?

Simplification for grouping of contracts

Where calculations need to be performed on a policy level under new run orders, that certain actuarial models could fall short very quickly when large policy counts are involved. However, to combat this low level of granularity, insurer’s may group policies together where reasonable information exists to motivate this.

Can I group policies together based on reasonable and available information to prevent contract by contract calculations being necessary? Should I start producing information that can inform this process?

Should I buy or should I build upgraded actuarial models?

More and more insurers are taking this opportunity to improve their actuarial modelling space and to buy off the shelf solutions from vendors who have catered for actuarial calculations under IFRS17. These sometimes also come with add-ons, for e.g. grouping policies by profitability as discussed above or aggregating projected cash flows over different levels, which you could leverage to speed up your implementation.

Some insurers prefer to buy new modelling software, whereas others have the experts inhouse and prefer to build their own solutions. Each comes with pro’s and con’s and this is a decision each insurer should investigate in great detail. Do you want to free up your scarce resources to focus on more important tasks (based on what you define to be important), or do you want them focusing on the actuarial models?

Given my available resources, flexibility in actuarial modelling software and speed at which I'd like to complete the build of my IFRS17 actuarial models, is it better to buy or to build IFRS17 actuarial models?

So where’s the line?

IFRS17 has shown that if you had slow, outdated models lacking flexibility, it’s probably time to upgrade to more robust and flexible modelling software. The amount of data required to be run through these models and output that needs to be generated has probably moved on to a place where we have to wave goodbye to our current models and welcome the newer and more robust actuarial modelling software out there. Actuaries and accountants all over are probably way more comfortable with their own models, which is understandable, but is it now perhaps the time to embrace change and move towards modernising your actuarial software? I think it’s time.

Need help on your IFRS17 implementation project? Then contact me on LinkedIn to see how we at Virtual Actuary can assist.

Did you have any comments? Please leave a comment and let me know what you think.

Did you like this blog? Then follow me and Virtual Actuary on LinkedIn for more IFRS17 blogs and keep an eye out for the next one.

What should I write about next? Let me know what you’d like me to write about next in the comments below.

Have any questions about Virtual Actuary? Visit our website at https://www.virtualactuary.com/ for more information.