Tree-based techniques have recently shown to be an interesting and effective tool to be used in establishing loss reserves on an individual basis. These new approaches to tackling the reserving task pose some challenges. After a brief literature review of the recent research works on the use of these strategies in a loss reserving context, we propose to investigate in detail methodologies based on survival analysis and the imputation of missing data to include open files within the loss reserving process. Through a simulation study, we compare their performance with some classical aggregate approaches, such as Mack’s Chain-Ladder and generalized linear models.
Pigeon et al. (Mon,) studied this question.