Stephan Bischofberger
Research Student
Cass Business School, Faculty of Actuarial Science and Insurance
Contact
- +44 (0)20 7040 5060
- stephan.bischofberger@cass.city.ac.uk
Visit Stephan Bischofberger
Bunhill Row
Postal address
106 Bunhill Row
London
EC1Y 8TZ
United Kingdom
About
Overview
Stephan is a PhD candidate in Actuarial Science at Cass Business School.
His research is mainly in statistical methodology for actuarial reserving but the forecasting problems on which he is working can also be found in many other fields including mortality, biostatistics and many other applications.
His work ranges from non-parametric statistics, survival analysis and computational statistics to the specific application in practice in reserving for general insurance.
At the moment, he is working on different continuous extensions of the chain ladder method involving kernel density and kernel hazard estimators. This involves various tasks from asymptotic theory to data application and simulation studies.
Interests:
Non-parametric statistics, local linear kernel estimation, smooth backfitting, survival analysis, counting processes, in-sample forecasting, general reserving, chain ladder method, operational time.
Qualifications
- MSc in Mathematics, Heidelberg University, Germany
- BSc in Mathematics, Heidelberg University, Germany
Languages
English, French and German.
Research
Title of thesis: Estimating operational time in a survival analysis setting and application in actuarial reserving
Sep 2016
Summary of research
The aim of my research is to add operational time, i.e., a certain dependence stucture, to the continuous chain ladder model. There is empirical evidence that such a generalisation of the continuous chain ladder model is appropriate and that independence should not be assumed. The goal is to derive full asymptotics for our new estimators and to compare it with another already existing one. Besides, we are interested in the results on small datasets as they can be found in practice.
Furthermore, I examine and compare the performance of simpler, multiplicative models for small samples sizes in simulations and on a dataset.
1st supervisor
- Professor Jens Perch, Professor of Actuarial Science
2nd supervisor
- Dr Munir Hiabu, Research Fellow
External supervisor
- Mammen, E. Heidelberg.
Publications
Journal article
- Bischofberger, S.M., Hiabu, M., Mammen, E. and Nielsen, J.P. (2019). A comparison of in-sample forecasting methods. Computational Statistics and Data Analysis, 137, pp. 133–154. doi:10.1016/j.csda.2019.02.009.
Working papers (2)
- Bischofberger, S., Hiabu, M., Mammen, E. and Nielsen, J.P. Smooth backfitting for additively structured hazard rates.
- Bischofberger, S., Hiabu, M. and Isakson, A. Continuous chain-ladder with paid data.