Seminars and events

Faculty of Actuarial Science and Insurance Research Seminars

Academic Year 2019/2020.

If you wish to attend a seminar, please book, using the link below the Seminar.  Tea and other light refreshments will be available 15 minutes before the talk begins in the milling area outside conference room.

The FASI seminars are recognised by the Institute and Faculty of Actuaries as providing 1 hour of continuous professional development (CPD) training.

If you would like to be added to the seminar electronic mailing list, please send an e-mail stating so, containing your name to

9th October 2019 - Hirbod ASSA

Dr Hirbod Assa
Head of Consultancy in Financial Math, University of Liverpool

No Hedging Market Consistent Valuation


The existing literature on market-consistent valuation (MCV) always assumes that the liquid assets cannot change the market valuation of a risky portfolio. In other words, an MCV needs also to be a hedging consistent valuation (or HCV). In this talk, we will discuss how HCV can be a restrictive condition when it comes to MCV. Then we consider a pure MCV, or MCV without hedging, and discuss how one can relate it to market sub-consistent valuation (MSCV). The discussions will follow by proposing methods for real word implement.


Dr Hirbod Assa joined the Department of Mathematics at the University of Liverpool in September 2013 as an assistant professor (Lecturer in the UK system). He now serves as the director of MSc program in financial mathematics and also as Head of Consultancy in the Department of Mathematics.  He earned his first PhD in mathematical finance from the University of Montreal, working on the application of risk measures in finance. In 2013 Dr Assa completed another PhD in economics at Concordia University, where he was awarded the Balvir Singh medal for outstanding achievements in my PhD thesis. He has experience of teaching courses in economics as well as in mathematics.   His background in two disciplines gives him  the ability to model real world problems with mathematical tools. His research interests cover different areas including insurance, re-insurance, asset pricing, agricultural insurance and game theory. Dr Assa is the founder and chair of Financial and Actuarial Mathematics group in Iran (FINACT-IRAN). He has chaired the  FINACT-IRAN conference in 2014, 2015, 2016 and 2017, and also organised few other workshops for this group.  He has been on the organizing committees of several events, including, , Quantitative Finance and Risk Analysis (2015, 2016) and Insurance: Mathematics and Economics (IME, 2015). Since September 2014  he has  become a member of the advisory board of Agricultural Finance Review

23rd October 2019 - David Atance del Olmo

Method for Forecasting Mortality based on Key Rates

Abstract: In this paper, we develop a model to construct dynamic life tables based on the idea that the behavior of whole life table can be explained by a reduced number of fac-tors. In this case these factors are identified with some mortality rates at specific ages. These key mortality rates and model parameters estimates are obtained applying a maximum likelihood criteria under the hypothesis of a binomial distribution of the number of deaths. In this paper, we develop the single factor version of the model which is implemented to the male and female populations of France and Spain.  The model is compared with a set of alternative well-known life tables models. To test the forecasting ability of the model we apply a battery of tests using out of sample data. Despite its simplicity, the outcomes indicate that this model it is not outperformed by other more complex mortality models. Another important advantage of this model is that can it be easily implemented to address some longevity risk linked problems in the context of Solvency II.

Keywords: Key mortality rate; Forecasting; Mortality modelling; Demography.


Cairns, A. J., Blake, D., Dowd, K., Coughlan, G. D., Epstein, D., Ong, A., and Balevich, I. (2009). A quantitative comparison of stochastic mortality models using data from England and Wales and the United States. North American Actuarial Journal, 13(1):1-35.

Debón, A., Montes, F., and Puig, F. (2008). Modelling and forecasting mortality in Spain. European Journal of Operational Research, 189(3):624-637.

Elton, E. J., Gruber, M. J., and Michaely, R. (1990). The structure of spot rates and immunization. The Journal of Finance, 45(2):629-642.

Haberman, S. and Renshaw, A. (2011). A comparative study of parametric mortality projection models. Insurance: Mathematics and Economics, 48(1):35-55.

Human Mortality Database (2019). University of california, berkeley (usa), and max planck institute for demographic research (germany).

Lee, R. D. and Carter, L. R. (1992). Modeling and forecasting us mortality. Journal of the American Statistical Association, 87(419):659-671.

Plat, R. (2009). On stochastic mortality modeling. Insurance: Mathematics and Economics, 45(3):393-404.

Renshaw, A. E. and Haberman, S. (2006). A cohort-based extension to the Lee-Carter model for mortality reduction factors. Insurance: Mathematics and Economics, 38(3):556-570.

Villegas, A. M., Kaishev, V. K., and Millossovich, P. (2018). StMoMo: An R package for Stochastic Mortality modeling. Journal of Statistical Software, 84(3):1-38.

Biography: David Atance is a PhD candidate in Economy and business management at Universidad de Alcala, Alcalá de Henares, Spain in his final year.   He has a Bachelor of Business Administration and Management and a Master of Actuarial Science and Finance in the University of Alcala (UAH).  His research interest focus on studying the evolution of mortality, prediction, forecasting and fitting different mortality model.   His current projects involve developing a model to construct dynamic life tables based on the idea that behaviour of whole life table can be explained by a reduced number of factor.

6th November 2019 - Jennifer ALONSO-GARCIA

Utility indifference pricing of a coupon-yielding bond.


When valuing claims on assets or indices which are not fully hedgeable, well-known option pricing expressions are no longer valid (see e.g. \cite{amin,black}) and incomplete market techniques need to be used. We use the theory of utility indifference pricing to derive a general framework to price claims on securities which are not traded \cite{musiela,young}. The utility indifferent price is the one that makes the issuer indifferent between issuing the claim, which involves receiving a premium and paying cash-flows throughout the duration of the contract and at maturity, and not issuing the claim. The strength of this technique is that it can incorporate the risk appetite of the issuer in the price and that it provides closed-form solutions when individuals have exponential preferences. We calculate the price for a security which is partially correlated with the financial markets and which makes regular coupon payments throughout the duration of the contract. Contrary to most frameworks found in the literature, we generalize the two-step pricing procedure and incorporate intermediate payments. The framework can be applied to the valuation of over-the-counter securities. In the specific insurance context, it can be used to price catastrophic-linked bonds or longevity bonds.


Jennifer Alonso García joined the Department of Economics, Econometrics and Finance (EEF) as an Assistant Professor at the University of Groningen in July 2018. Previously she was a Senior Research Associate at CEPAR, where she is now Associate Investigator. She completed her studies in mathematics in Spain and Germany, and received her PhD from the Université Catholique de Louvain in Belgium in 2015. She is a IABE Qualified Actuary of the Belgian Institute of Actuaries. She has also worked in the industry as a Risk advisor in the area or Solvency II and MCEV.

Her research combines the areas of actuarial science, household, pension and quantitative finance to study the design, risk-sharing and financing of funded and pay-as-you-go retirement income schemes. Her past, current and future research projects are all developed around the following overarching question: “How can we develop sound retirement income schemes that are fiscally sustainable and attractive for participants in an ageing environment?” During her PhD she studied both the fiscal sustainability and adequacy of pay-as-you-go financed defined contribution public pension schemes.

​Jennifer is currently involved in research projects on the financial decision making of households during retirement, life expectancy inequality and design and risk management of equity-linked retirement income products.  Her research has been published in leading international journals, including European Journal of Finance, Quantitative Finance, Insurance: Mathematics and Economics, Scandinavian Actuarial Journal and the ASTIN Bulletin and has been awarded the ICA 2018 Best Paper Award.

20th November 2019 - Anna Maria GAMBARO

Time Consistent Optimal Asset Allocation for Life Insurance Funds


In this work, we propose a dynamic consistent optimization problem for a portfolio of life insurance policies, in the Solvency II directive framework. In [Asanga et al., 2014], the authors use Solvency indicators to find the optimal asset allocation of non-life insurance funds, minimising the Solvency Capital Requirement (SCR), in a one-period model. [Christiansen and Niemeyer, 2014] propose a dynamic formulation of the SCR, using the dynamic value at risk (VaR). However, the dynamic-VaR is not a time consistent risk measure (see for instance, [Acciaio and Penner, 2010]). For non-life insurance funds and in case of a single liability cash-flow at maturity, [Devolder and Lebégue, 2017] analyse the time-consistent dynamic formulation of the SCR using the iterated-VaR and the iterated conditional tail expectation.  However, the iterated formulation of the SCR has some important drawbacks. [Devolder and Lebégue, 2017] show that using iterated risk measures, the SCR becomes quite expensive for long term liabilities and it may explode in some circumstances. Moreover, the iterated-SCR is not compliant with the Solvency II directive, in fact, it does not answer to the regulator request: how much is the capital to be held by insurance to meet his obligations over the following year?  We extend the literature in various directions. Firstly, we consider the framework of life insurance liabilities with multiple cash-flows. Secondly, starting from the static definition of the SCR in [Christiansen and Niemeyer, 2014], we propose a dynamic version of the SCR, that is time consistent, in agreement with the regulators directive and that encompass the drawbacks of the iterated formulation. Moreover, following the work of [Shapiro, 2009] on dynamic risk averse stochastic programming problems, we formulate a time consistent asset allocation problem based on the SCR minimization.

Finally, we apply the optimization problem to the case study of with-profit life insurance funds, which is analysed in [Gambaro et al., 2018] from the market-consistent valuation perspective.


Anna Maria Gambaro is an assistant professor in mathematics for economy, Finance and insurance at Università del Piemonte Orientale, Novara, Italy. Her research interests focus on pricing and risk management of financial and insurance products. Her recent research contributions have appeared in Quantitative Finance and Insurance: Mathematics and Economics. She obtained a Ph.D in Mathematical Finance in 2017 from Università degli studi di Milano Bicocca, faculty of Statistic and Quantitative Methods. Previously, she obtained an MSc and a BSc in Physics from Università degli Studi di Milano. From 2017 to 2018, she was granted a postdoctoral scholarship from Università del Piemonte Orientale, co-financed by Deloitte Consulting, Finance & Risk, under the supervision of Prof. Gianluca Fusai.

11th December 2019 - Michael SCHOLZ

Forecasting benchmarks of long-term stock returns via machine learning

Recent advances in pension product development seem to favour alternatives to the risk free asset often used in the financial theory as a performance standard for measuring the value generated by an investment or a reference point for determining the value of a financial instrument. To this end, we apply the simplest machine learning technique, namely, a fully nonparametric smoother with the covariates and the smoothing parameter chosen by cross-validation to forecast stock returns in excess of different benchmarks, including the short-term interest rate, long-term interest rate, earnings-by-price ratio, and the inflation. We find that, net-of-inflation, the combined earnings-by-price and long-short rate spread form our best-performing two-dimensional set of predictors for future annual stock returns. This is a crucial conclusion for actuarial applications that aim to provide real-income forecasts for pensioners.
Michael Scholz is an Assistant Professor at the Institute of Economics at the University of Graz. He graduated in mathematical statistics at TU Dresden and obtained his PhD in econometrics from the University of Göttingen. In 2017, he worked at TU Dortmund as a visiting Professor in Statistics. He is a specialist in non- and semi-parametric methods and computational statistics. He was and is currently engaged in a number of interdisciplinary projects working together with (development) economists, actuaries, statisticians, econometricians, or geographers. He published in a variety of top-ranked scientific journals and was recently awarded the John W. Kendrick Prize of the International Association for Research in

11th December 2019 - Stefan SPERLICH

Market-timing in Practise


Our interest in this paper is focused on actuarial models of long-term savings, and on potential econometric improvements of such models. The objective, in a nutshell, is to have a feasible method for a market timing strategy including comparisons of assets or portfolios. We will first revisit the utility function based investment approach that leads to a conditional mean-variance-ratio based strategy. While the general strategy, the setadaptive. The reason for this is our target of providing a Market time strategy that is feasible using only the limited information given in practise to e.g. insurance companies.


Stefan Sperlich made his diploma in mathematics at the University of Göttingen and holds a PhD in economics from the Humboldt University of Berlin. From 1998 to 2006 he was Professor for statistics at the University Carlos III de Madrid, from 2006 to 2010 chair of econometrics at the University of Göttingen, and is since 2010 professor for statistics and econometrics at the University of Geneva. His research interests are ranging from nonparametric statistics over small area statistics to empirical economics, in particular impact evaluation methods and empirical finance. He is cofounder of the research center 'Poverty, Equity and Growth in Developing Countries’ at the University of Göttingen, and is research fellow at the Center for Evaluation and Development (Mannheim, Germany). He published in various top ranked scientific journals of different fields and was awarded with the Koopmans econometric theory prize (among other prizes).

29th January 2020 - Xavier MILHAUD

Bias correction in machine learning for microlevel reserving

Register your attenance here.


In non-life insurance, business sustainability requires accurate and robust predictions of reserves related to unpaid claims. To this aim, two different approaches have historically been developed: aggregated loss triangles and individual claim reserving. The former has reached operational great success in the past decades, whereas the use of the latter still remains limited. Through an illustrative example and introducing an appropriate tree-based algorithm, we show that individual claim reserving can be promising, especially in the context of long-term risks.

Related paper:


Xavier Milhaud holds a Ph.D. in Applied Mathematics and Statistics from the University of Lyon, France. He is a fully-qualified actuary, member of the International Actuarial Association and worked for 4 years in AXA insurance company. He is currently Assistant Professor at ISFA, and his research interests focus on segmentation models and machine learning applied to actuarial sciences; with​ main applications on reserving and pricing.

More detailed information:

12th February 2020 - Yahia SALHI

Robust detection of abrupt changes in Poisson processes with application to the surveillance of biometric assumptions

Register your attendance here.


We consider the minimax quickest detection problem of an unobservable time of proportional change in the intensity of a doubly-stochastic Poisson process. We seek a stopping rule that minimizes the robust Lorden criterion, formulated in terms of the number of events until detection, both for the worst-case delay and the false alarm constraint. This problem, introduced by Page [Biometrika 41 (1954) 100–115], has received more attention in the continuous path framework (for Wiener processes) than for point processes, where optimality results only concern the Bayesian framework [In Advances in Finance and Stochastics (2002) 295–312, Springer, Berlin]. In this work, we prove the CUSUM optimality conjectured but not solved for the Poisson case of the CUSUM strategy in the general setting of the stochastic intensity framework. Applications to the surveillance of biometric assumptions in life insurance is discussed using real world datasets.


Yahia Salhi holds a PhD in applied mathematics from the University of Lyon, a MSc in actuarial science and finance from ISFA, and an engineering diploma from the Ecole des Mines. He is assitant professor at ISFA Graduate School of Actuarial Studies and associate researcher at the BNP Paribas Cardif's "Data Analytics & Models in Insurance's Chair" (DAMI). Yahia's main research interests include detection of abrupt changes, longevity and mortality modelling, pricing and management as well as surrender risk modelling and mathematical aspects of impairment of financial assets under IFRS regulations. Yahia lectures on actuarial and financial mathematics in various universities and actuarial programs: Saint-Joseph University (Lebanon), Université Internationale de Rabat (Morocco), Université Cheikh Anta Diop (Sénégal), Université Paris Dauphine (Tunisia) and at ISFA (France) among others.

26th Febuary 2020 - Kjersti AAS

Dependence modelling Via Vine Copulas

Register your place online here 

Vine copulas have proven to be flexible dependence models, which are able to model tail dependence pattern as they occur in financial and insurance data. The models power is driven by the ability to construct a d- dimensional dependence model from a collection of bivariate models through appropriate conditioning. This pair copula construction allows to work with high dimensional data. I will discuss the basic construction principle and give a review of its applications to finance.


Kjersti Aas is Assistant Research Director at NR, where she is also Research Leader for the research area Finance, insurance and commodity markets. She received her M.Sc. in Industrial Mathematics from The Norwegian Institute of Technology (NTH) in 1990 and her Dr.Philos. in 2008 from the Department of Mathematical Sciences, Norwegian Institute of Science and Technology (NTNU). She started as a research scientist at NR in 1991. Until 2000 her research was mostly within image analysis and pattern recognition. Since then her research has been focused on quantitative finance. Kjersti had a 20% position as adjunct professor at the Department of Mathematics, University of Bergen from 2011—2017, was key innovator in the research based innovation centre Statistics for Innovation 2007—2014, and is co-director for the research-based innovation centre BigInsight from 2015—2023. Kjersti has been project leader for more than 150 academic and industrial NR-projects during 1991—2019. She has published more than 25 peer-reviewed papers, and supervised 5 PhD candidates and 10 master students.

11th March 2020 - Mario WUTHRICH

From Generalized Linear Models to Neural Networks, and Back



We present how classical generalized linear models can be enhanced by neural network features. On the way there, we highlight the traps and pitfalls that need to be considered to get good statistical models. This includes the non-uniqueness of "sufficiently good" regression models, the balance property, and representation learning. These considerations bring us back to the concepts of the good old generalized linear models.



Mario Wüthrich is Professor in the Department of Mathematics at ETH Zurich, and Honorary Visiting Professor at City, University of London (2011-2022), he has been Honorary Professor at University College London (2013-2019), and Adjunct Professor at University of Bologna (2014-2016). He holds a Ph.D. in Mathematics from ETH Zurich (1999). From 2000 to 2005, he held an actuarial position at Winterthur Insurance, Switzerland. He is fully Qualified Actuary SAA (2004), served on the board of the Swiss Association of Actuaries (2006-2018), and is Editor-in-Chief of ASTIN Bulletin.

25 March 2020 - Tim VERDONCK

Fraud detection using analytics

Register your attendance online here

Financial institutions increasingly rely on predictive machine learning models to detect fraudulent transactions. Two main challenges when building a supervised tool for fraud detection are the imbalance or skewness of the data and the various costs for different types of misclassification. We discuss techniques to solve the imbalance issue and present a cost-sensitive logistic regression algorithm.


Tim Verdonck is a professor Statistics and Data Science at the Department of Mathematics of the University of Antwerp (Belgium). He is affiliated to KU Leuven and has been an invited professor at the University of Bologna, teaching advanced non-life insurance in the Master of Quantitative Finance. He is chairholder of the BNP Paribas Fortis Chair on Fraud Analytics, the Allianz Chair on Prescriptive Business Analytics in Insurance and the BASF Chair on Robust Predictive Analytics. His research interests are in the development and application of robust statistical methods for financial, actuarial and economic data sets. He is associate editor of Statistics: A Journal of Theoretical and Applied Statistics (Taylor & Francis) and Computational Statistics & Data Analysis (Elsevier). Tim is co-organizer of the Data Science Meetups in Leuven and managing partner at Boltzmann (, a team of experts in machine learning that transform data into actionable insights.

20 May 2020 - Juan SABUCO & Thorsten HEINRICH

Further information to follow in due course.

27 May 2020 - Katrien ANTONIO

Further information to follow in due course.

03 June 2020 - Eva CANTONI

Further information to follow in due course.

10 June 2020 - Alexander J. MCNEIL

Modelling Volatility with v-transforms

Register your place online here

A new approach to the modelling of financial return series using a class of transforms for uniform random variables is proposed. V-transforms describe the relationship between quantiles of the return distribution and quantiles of a positive-valued function of the return which acts as a predictable volatility proxy variable, such as the squared or absolute return. They allow the construction and estimation of models that combine arbitrary marginal distributions with linear or non-linear time series models for the dynamics of the volatility proxy. The idea is illustrated using a transfomed Gaussian ARMA process for volatility, yielding the class of VT-ARMA models. These can replicate many of the stylized facts of financial return series and facilitate the calculation of marginal and conditional characteristics of the model including quantile measures of risk. Estimation of models can be carried out by adapting the exact maximum likelihood approach to the estimation of ARMA processes.


Alexander McNeil has been Professor of Actuarial Science at the University of York since September 2016. Educated at Imperial College London and Cambridge University, he was formerly Assistant Professor in the Department of Mathematics at ETH Zurich and Maxwell Professor of Mathematics in the Department of Actuarial Mathematics and Statistics at Heriot-Watt University. He founded and led the Scottish Financial Risk Academy (SFRA) between 2010 and 2016.  His research interests lie in the development of quantitative methodology for financial risk management and include models for market, credit and insurance risks, financial time series analysis, models for extreme risks and correlated risks and enterprise-wide models for solvency and capital adequacy. He has published papers in leading actuarial, statistics, econometrics and financial mathematics journals and is a regular speaker at international risk management conferences.

He is joint author, together with Rüdiger Frey and Paul Embrechts, of the book "Quantitative Risk Management: Concepts, Techniques and Tools", published by Princeton University Press (2005/2015). He is also an Honorary Fellow of the Institute and Faculty of Actuaries and a Corresponding Member of the Swiss Association of Actuaries

Seminars take place on Wednesdays 16:00 to 17:00 in Room 2005. Light refreshments available at 15:45 in seminar room. The seminars are open to everyone.

Please contact: for further information.