Abstract
Bootstrapping is often employed for quantifying the inherent variability of development triangle GLMs. While easy to implement, bootstrapping approaches frequently break down when dealing with actual data sets. Often this happens because linear rescaling leads to negative values in the resampled incremental development data. We introduce two computationally efficient methods for avoiding this pitfall: splitlinear rescaling and parametric resampling using a limited Pareto distribution. After describing the essential mathematical properties of the techniques, we present a performance comparison based on a VBA for Excel bootstrapping application. The VBA application is available on request from the author.
Keywords:
Volume
11
Issue
1
Page
60-73
Year
2017
Keywords
Bootstrapping and resampling methods, generalized linear modeling, efficient simulation, stochastic reserving, regression, predictive analytics
Categories
Financial and Statistical Methods
Statistical Models and Methods
Boot-Strapping and Resampling Methods
Financial and Statistical Methods
Statistical Models and Methods
Generalized Linear Modeling
Financial and Statistical Methods
Statistical Models and Methods
Regression
Actuarial Applications and Methodologies
Reserving
Publications
Variance