Stochastic Loss Reserving for General Insurance with Emphasis on Micro-Level

Insurance companies ýusually use the chain-ladder method in their financial statements to predict loss reservesý. ý The chain-ladder method is based on aggregated data and development years of claims in run-off trianglesý. ýThis triangle is a summary of an underlying data-set related to the development of individual claimsý. This paper used ýthe framework of Position Dependent Marked Poisson Processes and statistical tools for recurrent events in individual claims for a method of reserve as a micro-level stochastic loss reserveý. Detailed information concerning damage occurrence timeýs,ý loss reporting delay timesý, intervals between payments, the values of their paymentsý, ýand the final settlement times of claims are used to calculate the micro-level reservesý. To validate the new model, ýthe data-set from an Iranian insurance company is considered. Using this data-set and the simulation of micro-level stochastic loss reserving model, it was shown that the use of micro-level stochastic loss reserving model is a close estimate to the real value of the required loss reserve in the future years.