Actuarial Science and Insurance Education and events

EM for Multivariate Claim Counts: Varying Dispersion and Dependence

Abstract:
This talk presents a regression framework for multivariate claim frequencies that accounts for overdispersion arising from unobserved heterogeneity and for dependencies between claim types that can be positive or negative. The focus is on multivariate count models in which Poisson components are linked through continuous latent effects with continuous marginals combined via copulas. This structure allows flexible dependence modelling and remains identifiable under mild regularity conditions. Estimation is carried out using a Monte Carlo Expectation–Maximization algorithm, which treats the latent variables as missing data and enables maximum likelihood inference when the joint distribution is intractable. A case study on the Wisconsin Local Government Property Insurance Fund shows that the proposed approach captures dependence patterns well and improves predictive performance compared to existing benchmarks. Diagnostic analyses further support the adequacy of the fit. The results highlight the importance of allowing both dispersion and dependen

Biography:
George Tzougas is an Associate Professor in the Department of Actuarial Mathematics and Statistics at Heriot-Watt University in Edinburgh, UK, and serves as the Academic Director of the Scottish Financial Risk Academy (SFRA). His research lies at the intersection of applied and computational statistics and statistical machine learning, with applications in insurance and, more recently, in computational finance. His work has been published in leading journals, including the Journal of the Royal Statistical Society (Series A & C), Journal of Computational and Graphical Statistics, Insurance: Mathematics and Economics, North American Actuarial Journal, ASTIN Bulletin, Scandinavian Actuarial Journal, Annals of Actuarial Science, European Actuarial Journal and Statistical Inference for Stochastic Processes among others. George’s recent research focuses on developing statistical machine learning models to evaluate the impact of climate hazards on non-life insurance portfolios and to advance the study of green finance. As Academic Director of the SFRA, he collaborates closely with industry partners to promote expertise in climate-related risk and has organized numerous events at Panmure House in Edinburgh. In 2026, he will serve as lead organizer of the international workshop “AI in Risk Assessment and Mitigation”, to be held in Edinburgh. The event aims to foster collaboration between academics and practitioners on how AI, statistics, and related methods can be applied to understand and manage societal risks, particularly those arising from climate change. He has received several distinctions for his research, including two Best Paper Awards from the Institute and Faculty of Actuaries (IFoA, 2021) and the SCOR Actuarial Award (2023).

Register here

Phased drawdown retirement plans and the gender pension gap

Abstract:
The progressive move from DB to DC pension plans is transferring from the state and employer to individuals the responsibility to manage the financing, investing, and spending of their retirement income. In countries as the UK and the US, financial institutions offer phased drawdown retirement plans which are popular among retirees to manage their pension pot although these products carry the risk of running out of funds. It is left up to the retiree to decide how to incorporate into their retirement plan their consumption needs, bequest preferences, investment preferences and, importantly, longevity risk. I investigate the effect on the retiree’s financial welfare of different investment strategies, consumption preferences and longevity risk.

The most important results include: phased drawdown retirement plans carry a significant probability of running out of funds; the probability of a female running out of funds can be 10% higher than that of a male; to keep the probability of running out of funds at an equal level a female will have a retirement income significantly lower than a male throughout all retirement; a gender pension gap is inherent to phased drawdown plans.

Keywords: Personal finance, retirement income plans, gender pension gap, longevity risk.

Biography:
Alexandra Dias is a Professor at the University of York who specialises in actuarial science and financial risk management. She received her PhD from the Eidgenössische Technische Hochschule Zürich (ETH Zurich) in Switzerland.

Alexandra has published research articles in finance, actuarial science, and mathematical statistics. Her current research is concerned with actuarial and economic aspects of pensions, including: retirement income strategies; consequences from replacing DB with DC pension schemes; inequalities in retirement income; and financial literacy.

Alexandra is a member of the Pensions Gap Working Party from the Institute and Faculty of Actuaries.

Register here

The short-term association between environmental variables and mortality: evidence from Europe.

Abstract:
Using fine-grained, publicly available data, this article studies the short-term association between environmental factors, i.e. weather and air pollution characteristics, and weekly mortality rates in small European regions. Hereto, we develop a mortality modelling framework where a baseline model captures a region-specific, seasonal trend observed within the historical weekly mortality rates. Using a machine learning algorithm, we then explain deviations from this baseline using features constructed from environmental data that capture anomalies and extreme events. We illustrate our proposed modelling framework through a case study on more than 550 NUTS 3 regions (Nomenclature of Territorial Units for Statistics, level 3) in 20 European countries. We show that temperature-related features are most influential in explaining mortality deviations from the baseline over short time periods. Furthermore, we find that environmental features prove particularly beneficial in southern regions for explaining elevated levels of mortality, and we observe evidence of a harvesting effect related to heat waves.

Biography:
Torsten Kleinow is Professor for Longevity Risk at the University of Amsterdam. His primary research interests include mortality modelling and the analysis of socio-economic disparities in life expectancy. He has published in international peer-reviewed journals and contributed to several international research projects on mortality and longevity.

Register here

A Utility-Based Life-Cycle Assessment of Reverse Mortgages in Retirement

Abstract
The financial well-being of older individuals is influenced by the stochastic nature of longevity and by changes in health status. Rising life expectancy in industrialized economies, coupled with the shortcomings of pension schemes and the burden of medical expenditures, underscores the need for suitable instruments to support retirees. Reverse mortgage (RM) contracts allow older homeowners to extract housing equity while retaining occupancy rights, with the outstanding balance—upon death—either repaid by the heirs or satisfied through the transfer of the property. Such arrangements may therefore represent a viable means of meeting consumption and care needs during retirement. We describe the decision-making problem faced by a homeowner approaching old age who must assess whether entering an RM contract is advantageous. To analyze how RMs a ect the consumption choices of older adults, we  develop a life-cycle model tailored to “house-rich, cash-poor” individuals, whose main asset is real estate and who must cope with uncertainty regarding lifespan, prospective health  conditions, and the associated costs of medical care and home maintenance. The lifetime utility framework incorporates consumption preferences, survival uncertainty, and bequest
motives, thereby capturing how emotional and financial attachment to housing influences retirement planning.

Biography
Gabriella Piscopo is Full Professor of Mathematical Methods for Economics, Finance, and Actuarial Sciences at the University of Naples Federico II. Her research focuses on actuarial mathematics, stochastic modeling, mortality and longevity analytics, and quantitative methods for insurance and financial risk assessment. She has authored numerous scientific publications on these fields, contributing to the advancement of stochastic approaches to risk theory, survival modeling, and data-driven techniques in actuarial and financial applications.

Register here.

Transformed gradient boosting and exponential dispersion models

Abstract:
Gradient boosting based models, including gradient boosting machines (GBMs), have proven to be highly competitive in predictive modelling. Since the introduction of GBMs in the early 2000’s, a large number of versions of gradient boosting methods have been developed. These include extensions such as XGBoost, NGBoost and lightGBM. Apart from going through the basic ideas underpinning GBMs, I will introduce a general transformed gradient boosting (TGB) method that can be used to provide theoretical guarantees of loss improvement for different versions of GBMs. In particular, this approach is used to motivate adjustments to NGBoost that improve the theoretical properties of  the algorithm, which also makes it more closely related to XGBoost. Further, in many actuarial applications the underlying stochastic model is a member of the exponential dispersion family (EDFs). When considering loss functions motivated by EDFs, it turns out that the TGB method can be significantly simplified, providing guidance on how to fit multi-parametric EDFs and how to include duration weights in a consistent manner.

This presentation is based on joint work with Ł. Delong, T. Nazar, and H. Zakrisson.

Register here

On optimal additional voluntary contributions (AVCs)
in DC pension schemes

Abstract:
In defined contribution pension schemes the member bears the investment risk and her main concern is to obtain an inadequate fund at retirement. To address inadequacy risk,  flexibility is often given to the member to pay additional voluntary contributions (AVCs)  into the fund. In many countries the AVC schemes allow members of the workplace pension  plan to increase the amount of retirement benefits by paying extra contributions.

In this talk, we will show how to choose optimally the additional voluntary contributions in DC pension schemes. This talk is based on two papers.
In the first paper,  in a simple Black and Scholes financial market, we define a target-based  optimization problem where the member of an AVC scheme can choose at any time the investment strategy and the additional voluntary contributions to the fund. In setting the problem,  the member faces a trade-off between the importance given to the stability of payments during the accumulation phase and the achievement of the desired annuity at retirement. We  derive closed-form solutions via dynamic programming and prove that (i) the optimal fund never reaches the target final fund, (ii) the optimal amount invested in the risky asset is positive, and (iii) the optimal AVC is higher than the target one. We run numerical simulations to allow for different member’s preferences, and perform sensitivity analyses to assess the  controls’ robustness. The second paper  extends the first one by considering an optimization problem for optimal investment and additional contribution in a more sophisticated financial-labour market with stochastic interest rate and stochastic salary. The optimal policies are provided in closedform and numerical simulations show the behaviour of the optimal policies over time and the impact of the stochastic salary on the optimal policies.

Biography:
Elena Vigna is full professor in Mathematical Methods for Economics and Actuarial and Financial Sciences at Università di Torino, Italy, and Affiliate of Collegio Carlo Alberto. From 2016 to 2022 she was Director of the Master Degree in Quantitative Finance and Insurance of Università di Torino. She holds a PhD in Matematica per le Decisioni Economiche from University of Trieste, and a Postgraduate Certificate in Actuarial Management from the City University, London. Her main research interests lie in the Insurance field, ranging from optimal decision making in defined contribution pension schemes to stochastic mortality modelling. Some recent research focuses on modelling and treatment of time (in)consistent dynamic optimization problems. She has several publications on the leading journals in Actuarial Science, and has published also on some high-impact journals in Mathematical Finance.

Register here

Robust distortion risk metrics and portfolio optimization

Abstract:
We establish sharp upper and lower bounds for distortion risk metrics under distributional uncertainty. The uncertainty sets are  characterized by four key features of the underlying  distribution: mean, variance, unimodality, and Wasserstein distance to a reference distribution. We first examine a broad class of distortion functions, assuming only finite variation and imposing neither continuity nor monotonicity. This class includes important examples such as the Gini deviation, the mean–median deviation, and inter-quantile differences. When the uncertainty set is defined by a fixed mean, variance, and Wasserstein distance, we derive the worst- and best-case values of the distortion risk metric and identify the corresponding extremal distributions.

We then impose an additional unimodality constraint.   In this case, for absolutely continuous distortion functions, we again characterize the worst- and best-case values and explicitly determine the optimal distributions attaining these bounds.

Finally, we illustrate the practical relevance of our theoretical results in the context of robust portfolio optimization.  (This talk is based on a joint work with Steven Vanduffel and Yi Xia).

Biography:
Dr. Peng Liu has been a Lecturer at the University of Essex since 2020. He obtained his PhD in Probability and Statistics from Nankai University in 2015. He subsequently held postdoctoral positions at the University of Lausanne and the University of Waterloo. His research primarily focuses on quantitative risk management, actuarial science, financial mathematics, and extreme value theory. His work has been published in leading journals in the field, including Mathematics of Operations Research, Mathematical Finance, Finance and Stochastics, SIAM Journal on Financial Mathematics,  European Journal of Operational Research, Stochastic Processes and Their Applications, and Insurance: Mathematics and Economics.

Register here

Incorporating Heart Rate Variability into Health Insurance Pricing: Evidence from Simulation and Wearable-Derived Data

Abstract:
Heart rate variability (HRV), measured using wearable devices, is recognised as being associated with mortality risk. However, the implications of incorporating HRV into life-insurance pricing, particularly in the presence of information asymmetry between individuals and insurers, remain unclear. This study examines how HRV-based risk information may affect insurer performance under alternative pricing strategies using a controlled simulation framework. We generate a synthetic population calibrated to age- and sex-specific baseline mortality rates and introduce an HRV-related multiplicative hazard component, with effect sizes varied across a plausible range informed by the clinical literature. Two insurers are considered: one that prices policies using traditional rating factors only, and another that incorporates HRV into premium calculations. Individuals are assumed to purchase insurance based on premium price, while only a proportion of the population is informed about their true HRV-adjusted risk. Both the strength of the HRV–mortality association and the proportion of informed individuals are varied, with each scenario repeated multiple times to account for stochastic variability.

Register here

Dependence Modeling in General Insurance using Local Gaussian Correlations and Hidden Markov Models

Abstract:
Pearson’s rho is a commonly used measure of dependence but can be misleading in heavy-tailed or nonlinear settings. This talk examines Local Gaussian Correlation (LGC) as an alternative, combining it with Hidden Markov Models (HMMs) to capture time-varying and regime-specific dependence structures in general insurance data. We propose a bootstrap test to assess differences in dependence across regimes and apply the framework to insurance claims and financial data. The results indicate structural breaks and varying tail dependencies over time. We also demonstrate how the model can be used to produce regime-aware risk measures such as Value-at-Risk (VaR) and Tail VaR, offering a more adaptive approach for risk assessment in insurance applications.

Biography:
Bård Støve is a Professor of Statistics at the Department of Mathematics, University of Bergen, Norway. He holds a PhD in Statistics from the University of Bergen and an MSc in Mathematics and Physics from the Norwegian University of Science and Technology. He is a qualified actuary and full member of the Norwegian Actuarial Association. His research focuses on the development of statistical methodology—particularly nonparametric methods, time series analysis, and dependence modeling—with applications in finance, insurance, climatology, and medical statistics. He has published in leading journals, including Statistical Science, Journal of Empirical Finance, Journal of the Royal Statistical Society, and Insurance: Mathematics and Economics. In addition to his academic role, he holds a part-time position as an actuarial consultant, contributing expertise in reserving, pricing, and regulatory compliance.

Register here