Can someone recommend a good paper to read on bootstrapping (chain ladder or ODP) - or perhaps just somewhere where I can get more info (such as a site online) I want to try do this in R - any advice if you know how or have attempted this in R before..?
To be honest, the answer to this depends a lot on what your aims are. The R 'chainladder' package already implements the bootstrap. If you simply want to learn the calculations involved without too much emphasis on theory perhaps it would be better to work from an example spreadsheet. I can probably dig one of these out if you like. Alternatively, if you actually want to understand the reasoning that leads to these calculations, have a look at England & Verrall's "Analytic and bootstrap estimates of prediction errors in claims reserving" (1999) along with the addendum (2002).
Hi td290 - I would really, really appreciate it if you could pass on that example spreadsheet! Thanks very much But ultimately I want to understand the reasoning behind what I'm doing as well, so thanks for that recommendation. I'm quite interested in this process and how and why it works
In the ChainLadder package in R, does anyone know the code to output the triangle(s) associated with a reserve at a certain percentile? So, for example, if the 75th percentile IBNR was found to be X, what triangle(s) were generated in the bootstrapping process to produce this result of X?