Regression approach to technical reserve calculation

Phuong Dao Thi Ha

Master's thesis

Abstract:

In this thesis, we consider several existing models in a claims reserving problem. We compare the triangle model and generalized linear models (GLM) through their predictions. Both the theory of GLM as well as reserving theory are discussed. The practical part of our work is focused on the application GLM techniques and comparison of these models on real data represented by a non-life insurance portfolio. Mean squared error of prediction (MSEP) and bootstrapping are used to compare these models. Our goal is to find the structure of the best model suited for the given dataset measured using quality assessments.

Abstract:

In this thesis, we consider several existing models in a claims reserving problem. We compare the triangle model and generalized linear models (GLM) through their predictions. Both the theory of GLM as well as reserving theory are discussed. The practical part of our work is focused on the application GLM techniques and comparison of these models on real data represented by a non-life insurance portfolio. Mean squared error of prediction (MSEP) and bootstrapping are used to compare these models. Our goal is to find the structure of the best model suited for the given dataset measured using quality assessments.

Keywords

claims reserving, over dispersion, non-life insurance, triangle model, bootstrapping, gamma, mean squared error, generalized models

Keywords

gamma, bootstrapping, generalized model, non-life insurance, triangle model, claims reserving, over dispersion, mean squared error

Language used: English

Date on which the thesis was submitted / produced: 4. 9. 2017

Thesis defence

• Date of defence: 11. 6. 2018
• Supervisor: Pavel Zimmermann
• Reader: Václav Sládek

Vysoká škola ekonomická v Praze

Master programme / field:
Kvantitativní metody v ekonomice / Quantitative Economic Analysis