Since the time of the MCEV models in the early 2000’s European life insurance companies have been facing a computational challenge: How to accurately project the entire portfolio consisting of thousands of policies with the existing IT capacities? The models become ever more complex, due to regulatory and internal company requirements. Moreover, the entire portfolio must be projected far into the future, in case of life-long contracts often 60 years or more. To make matter worse, there is no analytical expression for the best estimate of the portfolio value (own funds from shareholders’ perspective and best estimate liabilities from policyholders’ perspective). Therefore, companies have to perform thousands of probability weighted Monte Carlo simulations. All this taken together leads to exploding runtimes even on large, expensive in-house server farms or with an efficient usage of cloud technology.

Although the so-called Moore’s Law, according to which the computational capacity doubles every two years, has maintained its validity during the last two decades, this was not enough for the life insurers to be able to meet the deadlines and keep the IT costs under control. In most cases, the runtime of cash flow projection models greatly exceeds the available time for Solvency II or IFRS closure, if companies attempt to project each policy separately.

The solution the companies have developed over the years consists in identifying and grouping similar policies and projecting only one representative of each group. This clustering of policies is known as grouping or optimization of the portfolio.

In the web session we will show various methods for grouping of life insurance policies, starting from the more traditional approaches of non-negative least-squares optimization, over the methods of unsupervised machine learning like k-means clustering and supervised learning like neural networks, all the way to the most recent methods leveraging on the simplex methods from mathematical programming. This is an active research area and the latest successes in developing efficient methods without a need of manual intervention will also be presented. 

How these methods perform on different realistic portfolios of policies will be another important part of the web session.

Organised by the EAA - European Actuarial Academy GmbH.

The expected audience for this web session will be practitioners who work in the area of cash flow projection models. We deem it useful for a professional struggling to meet challenging Solvency II deadlines or a manager who is not sure how to deliver the granularity requirements of the IFRS 17 reporting and keep the accuracy of the results. Additionally, practitioners who are interested in ways how modern machine learning techniques can be successfully implemented in actuarial modeling may find the topic interesting.
Technical requirements and test session
Please check with your IT department if your firewall and computer settings support web session participation (the programme GoToTraining is used for the web session). Please also make sure that you are joining the web session with a stable internet connection.

On 9 June 2020, 10:00 - 10:30 CEST, there will be a test session offered to all registered participants to test the software.

We will give an overview over the methods used for the clustering of life insurance policies for the purpose of projecting porfolios for the risk-neutral valuation. This clustering, also called grouping or optimization of portfolios, is important, because the costs of cash flow projections of full insurance portfolios are prohibitively high. Depending on the homogeneity of a portfolio of policies, grouping can reduce the size of the portfolio by factors between 50 and 10,000. 

The language of the web session will be English.

Prof. Dr. Christian Weiß, Dean and Professor, Ruhr West University of Applied Sciences
In 2012, Christian finished his PhD at Goethe-University, Frankfurt. After a short period as postdoc, he moved to the financial sector working as an actuary in risk management for Generali Germany. In 2017, he became professor for mathematics at Ruhr West University and one year later dean of faculty. Among his main areas of research are Quasi-Monte Carlo methods and applications of machine learning to insurance mathematics. Besides his academic activities he works as a consultant for B&W Deloitte.

Dr. Zoran Nikolić, Partner, B&W Deloitte, Cologne
Zoran is responsible for the actuarial data science area within B&W Deloitte office in Germany and Austria. He has previously worked in the risk management and actuarial departments of Generali in Germany, where his focus were actuarial cash flow models, proxy models and stochastic valuation of life and health insurers. Zoran holds a PhD from University Göttingen and is also active as lecturer at the Mathematical Institute in Cologne and the German Actuarial Association.