MODELLING DEPENDENCIES USING GENERALIZED ESTIMATING EQUATION TO ESTIMATE CLAIM RESERVING
In a non-life insurance business, the outstanding claims reserve is one key factor which will determine the financial position of an insurance company. Two approaches may be used to predict the outstanding claims reserve: deterministic and stochastic. In this thesis, astochastic model, nam...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/21272 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | In a non-life insurance business, the outstanding claims reserve is one key factor which will determine the financial position of an insurance company. Two approaches may be used to predict the outstanding claims reserve: deterministic and stochastic. In this thesis, astochastic model, namely Generalized Estimating Equation(GEE), is used to predict the outstanding claims reserve for a long tail insurance business. The data for such business may be represented in a run off triangle format. In GEE, the data in one accident year is treated as a panel data. Every data in one panel is assumed to be correlated and the correlation matrix may be chosen from one of three correlation structures, namely: Autoregressive of order-1 (AR-1); Exchangeable and Independence. The suitability of the model chosen is determined by conducting diagnostics tests on the residuals. The most suitable model is chosen using the Quasi Information Criterion (QIC); the Correlation Information Criterion (CIC); and the Prediction Mean Square Error in GEE (PMSEG). The cluster bootstrap process is included to determine the confidence interval of the claims reserve. These methodologies are applied to a set of data obtained from State Farm Mutual Insurance’s company for accident years 1988-1997 in the National Association of Insurance Commissioners’s (NAIC) database. |
|---|