EAA Web Session: "Non-Life Reserving"

Einleitung/Dauer

As Insurance is being impacted by new and disruptive technology, how will actuarial reserving techniques be impacted? Whilst triangle methods have traditionally been seen as the key methods in the reserving area, this may be challenged in the future by more complex expectations, improving technology, modelling capabilities, etc.

Reserving practices are expected to keep changing, as the influences of big data and regulatory requirements continue to evolve.

The aim of this workshop is to

• Refresh classical actuarial/statistical techniques used in non-life reserving (Gaap and Solvency 2).
• Focus on some practical problems faced by reserving actuaries, by presenting practical examples.
• Introduce more advanced techniques used in non-life reserving in order to open new perspectives on more granular reserving calculation, case estimates, etc.

Preliminary Programme
Monday, 9 March 2026
Non-life reserving and risk management
e-learning prerequisites (~ 1h30): Non-life premiums and premium provisions; Claims and claims provision; Introduction to Solvency 2
09.00 – 12.30 (including a break of approx. 10 minutes) 
• Refresh on Gaap and Solvency 2 (S2) reserves
• Refresh on triangle reserving methods
• Standard Deviation: one-year and ultimate view with analytical formula
• Entire distribution estimate: bootstrap approach
• Premium & reserve risk in Solvency II: Standard Formula & Undertaking Specific Parameters
• BeGAAP & LAT (Liabilities adequacy test): margin management
• Example: Reserving with R using triangle data

Tuesday, 10 March 2026
Generalized linear models for non-life reserving
e-learning prerequisites (~ 30min): Generalized linear models
09.00 – 12.30 (including a break of approx. 10 minutes)
• Refresh on GLM models
• GLM models on triangle data: multi-triangles approach
• Best practices: outlier detection, data importance, smoothing, etc.
• GLM models for opening case reserves
• Example: Reserving with GLM in R using triangle data

Wednesday, 11 March 2026
Advanced methods of non-life reserving with R
e-learning prerequisite (~ 1h):

• Supervised Machine Learning (part 1: error measure, decision tree, bagging and random forest)
• Supervised Machine Learning (part 2: gradient boosted methods, neural networks, support vector machine)

09.00 – 12.30 (including a break of approx. 10 minutes)
• Available techniques for individual claims reserving
• Pros and cons of aggregate versus individual reserving 

All the above times are given in CET (Central European Time).

Vorgehensweise und Ziele

The seminar will alternate between methodological concepts and practical examples in order to ensure a comprehensive understanding of the techniques presented.

The instructor-led sessions will take place online over three days between 9:00 and 12:30.

The participants will be requested to look at 5 e-learning capsules (of around 30 minutes each):

• Non-life premium and premium provision
• Claims and claims provisions
• An introduction to Solvency II
• Supervised Machine Learning (part 1: error measure, decision tree, bagging, and random forest)
• Supervised Machine Learning (part 2: gradient boosted methods, neural networks, support vector machine)

The participants will be requested to watch 1 recorded session (of around 30 minutes):

• Generalized linear models

The e-learning capsules and recording contain pedagogical presentations of the concepts with examples. They can be followed by the participants whenever they want between the web sessions, as pre-requisites.

The examples are developed in R. We encourage participants to actively participate during the sessions.

Examples in e-learning capsules and during the sessions are based on our experience with Belgian companies but are not specific to Belgium and are rather applicable to most European countries.

Teilnehmer

This training course is designed for people with an actuarial background who wish to develop or refresh their knowledge in the field of non-life insurance reserving. A basic knowledge of the R programming language is strongly recommended in order to make the most of the lessons and examples.

Technical Requirements
Please check with your IT department if your firewall and computer settings support web session participation (the programme Zoom will be used for this online training). Please also make sure to join the web session with a stable internet connection.

Dozierende

Maxime Doumont
Maxime joined Reacfin in October 2024 as part of the Non-Life Center of Excellence. He holds a Bachelor’s and Master’s Degrees in Engineering (with a specialization in Mathematics)  as well as a Master’s Degree in Actuarial Sciences from the Catholic University of Louvain (UCLouvain, Louvain-la-Neuve). He is also a Junior Member of the Institute of Actuaries of Belgium (IA|BE). Before starting as a full-time employee, Maxime performed a 1-year traineeship at Reacfin, working on pricing under ethical constraints (interpretability, non-discrimination and fairness). During his work experience, he developed a valuable knowledge in Reserving (implementation of intra-year methodologies, stochastic reserving using GLM with an inflation component) and Pricing (implementation of a churn and customer lifetime value model).

Geoffrey Feraut
Geoffrey Feraut is a manager at Reacfin and is the current Head of the Non-Life Center of Excellence. He joined Reacfin in March 2020 after four years at Cigna Europe Risk Management Team. During his time with Cigna, Geoffrey built a good experience on Solvency II Pillar 1 and developed his technical skills in R. As a Reacfin consultant, he had the opportunity to value his Solvency II experience through a general support mission (Best Estimate and SCR calculation on a quarterly basis) as well as support for the IMAP of an internal Non-Life reinsurance entity. Geoffrey is also involved in projects related to IFRS17. Geoffrey holds a Bachelor’s and Master’s Degrees in Physics as well as a Master’s Degree in Actuarial Sciences from the University of Louvain (UCLouvain). Geoffrey is a qualified actuary of the Institute of Actuaries in Belgium (IA|BE).

Xavier Maréchal
Xavier is founder and CEO of Reacfin. Xavier is one of the co-authors of “Actuarial Modeling of Claim Counts: Risk Classification, Credibility and Bonus-Malus Systems” (Wiley, 2007). Xavier has obtained different academic degrees as MSc. Civil Engineer (Applied Mathematics), MSc. Actuarial Science and MSc. Management. Xavier has extensive experience in the actuarial field obtained during his 15 years as a principal consultant for many national and multinational insurance companies. He has gained complementary experience in various fields going from Non-Life ratemaking and provisioning to health modeling and ALM. After several years of intensive modeling activities in Health, Non-Life and ALM, Xavier works now mainly as reviewer, mentor for consultants, and trainer. He performed several validation assignments and holds the actuarial function for a health insurance company. Xavier is a qualified actuary of the Institute of Actuaries in Belgium (IA|BE) and is involved in the Data Science and AI workgroup of the IA|BE.

Sprache/Kurztitel

The language of the web session will be English.

CPD Credits
For this web session, the following CPD credits are available under the CPD scheme of the relevant national actuarial association:

• Austria: 10 points
• Belgium:  10 points
• Bulgaria:  12 points
• Croatia: individual accreditation
• Czechia:  10 hours
• Denmark: 9 credits
• Estonia:  10 hours
• Finland: 7 points
• France:  54 points
• Germany:  9 hours
• Greece: 13 points
• Hungary:  10 hours
• Iceland: 10 credits
• Ireland: 10 hours
• Italy: approx. 4 credits (GdLA individual accreditation)
• Latvia: 10 hours
• Lithuania: 10 hours
• Netherlands:  approx. 10 points (individual accreditation)
• Norway: 10 points
• Poland: 10 hours
• Portugal: 10 hours
• Serbia: 5 hours
• Slovakia:  8 points
• Slovenia:  50 points
• Spain: CAC: 10 hours, IAE: 10 hours
• Switzerland:  15 points
• USA: SOA (Section B): up to 12 hours

No responsibility is taken for the accuracy of this information.