The Convergence and Robustness of Cohort Extensions of Mortality Models

In short

For my bachelor, I wrote my thesis within the domain of Actuarial Sciences on the subject of mortality models. More specifically, on a specific cohort extension of the well-known Lee-Carter model. It is published through the MaRBLe program and available through a link at the bottom of this page under the resources section.

Background of the Bachelor

In the 2nd semester 3rd year of the Bachelor Econometrics & Operations Research, students are required to choose a minor (15 ECTS) and write a Bachelor thesis (15 ECTS). I opted for Actuarial Sciences, meaning a course heavily focused on mathematical computations covered in this book and a thesis in similar direction, both under supervision of Eric Beutner.

The Thesis

I wrote my thesis on the topic of Mortality Models, specifically models known as the Lee-Carter models. These models are used to predict the life expectancy of people over time, which might sound a little bit macabre at first but given modern societal concepts such as pensions and life insurances, it is of vital importance to be able to appropriately predict life expectancy. Not to predict when people will die, but how long they probably will live and what that financially means for our society. Anyway, to be more specific, I reconstructed and applied some of the mathematics behind Cohort extensions of the model in order to assess their statistical convergence and robustness. In other words, are the statistics behind the model converging towards correct statistical decisions even when the initial assumptions in the mathematical derivation are somewhat relaxed.

On MaRBLe and publishing

The Maastricht University School of Business and Economics faculty has a honours-program for ambitious bachelor students called MaRBLe (Maastricht Research Based Learning), for which students of Econometrics and Operations Research are automatically enrolled. To be honest, I am still not aware of what MaRBLe exactly entails, but I did come to learn that it is responsible for the yearly publishing of several MaRBLe research papers. The idea behind this is to give bachelor students an early publication in their carreer and potentially steer it towards more of an academic direction. Every year the 4 econometric disciplines are represented through one thesis, namely:

• Auctions and Electronic Markets
• Econometric Methods
• Advanced Algorithms
• Actuarial Mathematics

As such, I’m not sure whether the best theses are part of this publication and I would not dare to entertain that idea. I always had a lot of respect for the intelligence and academic abilities of my peers and I cannot really imagine to have written the best thesis. Perhaps me starting a Research Master and actively seeking out a more academic carreer at that time helped me but in the end the professors behind the minors decided.

Resources

Please find the thesis here: