Μέθοδοι τιμολόγησης προϊόντων γενικών ασφαλίσεων με βάση την Ευρωπαϊκή Οδηγία Φερεγγυτότητα ΙΙ
Methods of ratemaking for non-life insurance based on Solvency II
KeywordsΜέθοδοι τιμολόγησης ; Solvency II ; Γενικές ασφαλίσεις ; Μέθοδοι τιμολόγησης γενικών ασφαλίσεων ; Τιμολόγηση ασφαλίστρου ; GLM ; GLMM ; Κλάδος γενικών ασφαλίσεων ; Τιμολόγηση ; Εκτίμηση ασφαλίστρου ; Φερεγγυότητα II
Pricing in general insurance differs significantly from pricing of products/goods or other kinds of services. It is a more complicated procedure as it takes into consideration multiple parameters and future costs that need to be calculated to present prices. Moreover, the fact that each kind of insurance has to deal with a different kind of risk (Exposure) with disparate groups of insured makes the whole process of pricing a lot harder. In this thesis, a reference to the fundamental terms of insurance and more specifically of general insurance is used, in order to familiarize the reader with the pricing procedure. Afterwards, there is a reference to the Solvency II prudential framework and the requirements it establishes to insurance organizations regarding the economic capital required, in order to assure the organization’s solvency and to minimize the risk of bankruptcy. Moreover, the process of estimating the pure premium is analysed, while factoring in the expenses, the cost and the target-profit associated with the premium Despite the fact that general insurance consists of different kinds of insurance programs and therefore premiums, such as fire insurance, vehicle insurance, property insurance etc., the examples and the simulation are centered around vehicle insurance with an unchanged Exposure, in order to help the reader understand the terms used, as well as the comparison between the different linear models and distributions used. The simulation focuses on the estimation of Pure Premium. In the application of the simulation of linear models through the R project, we try not only to estimate the Pure Premium but to evaluate the credibility of the model used for the pure premium. This evaluation is achieved through the use of three measurements, first without random effects, then with the use of one random effect and last with two random effects. The goal of this evaluation is to test to what degree the evaluation process picks out the same model for frequency and severity as the one used in the simulation. This procedure leads to clear indications regarding the level of the model’s credibility.