Risk adjustment: Opportunities and challenges for non-life businesses

Risk adjustment: Opportunities and challenges for non-life businesses

By Kevin Poulard, Head of R&D – IFRS 17 Expert, addactis

Can you please tell the Risk Insights readers a little bit about yourself, your experiences and what your current professional focus is?

I am head of research and development at addactis® Software. This is a new role for me at the intersection of actuarial science and technology, providing technical vision among our software range from modeling life aspects to reserving. We are covering the whole range of (re)insurance needs: pricing, reserving, reporting and modeling.

As for providing support on the actuarial aspects, I have been involved in the challenging IFRS 17 developments to make our tools compliant with the standard. In fact, IFRS 17 has impacts on a wide variety of topics ranging from data management to core actuarial functions. For this reason, we have the possibility to interact with different tools that we deliver to provide a customizable, user-friendly framework for our clients where we focus on business-driven automation.

Previously, I have been working on addactis® Modeling; our actuarial modeling and risk management platform for both life and non-life (re)insurers. I have been working on different standard actuarial models such as the Economic Scenario Generator, Life cash flows and assets projections as well as Solvency II.

What, for you, are the benefits of attending a conference like the IFRS 17 Forum and what can attendees expect to learn from your session?

Attending the IFRS 17 forum is an undeniable asset to share our understanding of the IFRS 17 standard whilst confronting insurance companies’ practices.
From the granularity and associated data requirements to the various possibilities of calculating the Risk Adjustment, many questions arise around IFRS 17 implementations.

During my speech, I will try to explain challenges regarding the assessment of the Risk Adjustment for General Insurance. In fact, IFRS 17, as a principle-based standard, does not say much on the implementation aspect. However, it is the task of the actuaries to understand the core of this concept and how IFRS 17 expresses it. I will show how to integrate a granular calculation using different approaches and a traditional reserving framework, most commonly called run-off triangles. We will go through the IFRS 17 specifications regarding the Risk Adjustment; equivalent confidence level, granularity requirement, aggregation and reallocation.

Please could you provide an overview of the challenges and opportunities of the parametric approach?

There are two different methodologies; a simulated-based approach and a parametric approach.
The simulated-based approach requires to derive simulated expected cash flows, hence, stochastic reserving methods have to be used; either projecting claim by claim or using aggregated triangle models. This will bring challenges in understanding the factors that drive the distribution but enabling to aggregate distributions and evaluate the diversification in a consistent manner.

When deriving the Risk Adjustment, the parametric approach requires calibrating the distribution and its associated parameters. This could be seen as a Solvency II framework, less demanding regarding memory management and keeping the standard deterministic reserving method. However, estimating distribution moments could be difficult (especially for moments higher than order 3). Moreover, aggregation will be a real challenge when using other distributions than elliptical one. Therefore, explicit assumptions have to be considered.

An interesting aspect is that both methods could be used together depending not directly on the underlying insurance contract, but the nature of the claim. From what we have seen working with a major Pan-European insurance firm, the parametric approach could be well fitted when working on attritional claims. In fact, non-proportional reinsurance contracts held on those types of claims will not be triggered and gross to net ration on proportional reinsurance could be adequate used cautiously. When reserving large claims, non-proportional will be triggered and a simulated approach becomes inevitable to assess the risk under such claims.

How are the challenges in risk adjustment affecting re-insurance?

In order to understand challenges while computing the Risk Adjustment for reinsurance, the actuary has to understand the measurement of such contracts under the IFRS 17 standard. In fact, reinsurance contracts held have to be recognized, measured and disclosed separately from the gross figures. Moreover, the cash flows distribution has to reflect the risk of non-performance by the reinsurer.

Therefore, applying gross-to-net ratio for the whole reinsurance structure (proportional and non-proportional) seems inaccurate. Then, estimating risk relating to reinsurance contracts will be complex without the use of a simulated approach.

Finally and not to forget, insurance companies have to think about contract boundaries issues. From a reinsurance contracts point of view, it can cover insurance policies that have not been signed yet. At the same time, from a policy point of view, some future claims related to current policies would be covered by reinsurance treaties that have not been signed yet. It will imply mismatch between gross and net of reinsurance results because of the underlying measurement models and constraints brought by the standard.

BBA vs PAA: What should be taken into account when measuring the RA?

To sum up the standard, the Premium Allocation Approach is focusing on assessing the risk of the LFIC (Liability for Incurred Claims). Using this framework, the LFRC (Liability for Remaining Coverage) is approximated and an explicit calculation of the Risk Adjustment has not to be derived. On the contrary, using the Building Block Approach, an explicit Risk Adjustment has to be disclosed for both the LFIC and LFRC.

For General Insurance, what we have seen when assessing the standard during client implementation is the fact that in the current process, LFRC is roughly calculated as the PAA was used (unearned premium). Even if General Insurance expects to use PAA for most of their businesses, they still have to prove the validity and the non-onerous characteristic of their long-tail businesses. Therefore, computing the Risk Adjustment for the LFRC is a challenge regarding the current reserving processes. In fact, a distribution has to be derived and working on an Accident Year basis will reveal the lack of data in estimating this quantity. From expert judgments to reserving process improvement, the calculation of the Risk Adjustment for the LFRC requires to estimate at least the associated cash flows and its underlying risk.

How do you see the impact of IFRS 17 evolving over the next 6-12 months?

IFRS 17 requires rethinking the risk management process whilst leveraging Solvency II investments. For General Insurance, it will imply to automate the reserving process as much as possible but also keeping a general reserving framework to avoid misunderstood gaps between Solvency II and IFRS 17 balance sheets. Regarding the implementation process, it is an on-going process but at different stages depending of the size of the company. Best practices will arise over time, and, regarding this topic, the International Actuarial Association has recently published a Risk Adjustment monograph. However, time is crucial and costly regarding the deadline even if the IASB recently voted in favor of a one-year delay.

You may also be interested in…

Sign up for your free account: