The need for diagnostic assessment of bootstrap predictive models

Glen Barnett and Ben Zehnwirth

Contents:

• The need for diagnostic assessment of bootstrap predictive models
• A basic bootstrap introduction
• Diagnostic displays for a bootstrapped chain ladder
• Assessing bootstrap predictive distributions
• Some other considerations
• Conclusions
• References
• Appendices

The following set of pages are available as a PDF document here.

Appendices

Appendix A

Incurred Loss Array for the Mack (RAA AFG) data

Appendix B

ABC - Incremental paid losses triangle and exposures

Note that the exposures were not used in the chain ladder fits described here. (Since there is already a parameter to each accident year, there is little point in any case.)

Appendix C

LR high data, incremental payments, development years 0-8

LR high data, incremental payments, development years 9-17

Appendix D

Taylor-Ashe data. Incremental payments

Taylor-Ashe data. Number of claims finalized

Note that Taylor and Ashe (1983) give the data as payments per claim finalized and number of claims finalized. The original payments have been reconstructed. Note also that the number of claims finalized have not been used in the analysis here, the table is given for the sake of completeness.