| Abstract: | In this paper, we consider the question of predicting the final amount of a claim and its distribution from micro-level data. A copula model is used to describe the dependence between the amount of a claim and its duration (that is the time between its occurrence and its closure). Due to the presence of censoring, we adapt classical methodologies using a weighting scheme that corrects the bias caused by this incompleteness in the data. Theoretical results and simulation support the validity of the procedure. A real case coming from medical malpractice claims is presented. |
