Bootstrapping and Mack assumptions

[code9063]

Hi,

One of the bootstrapping method assumption is "incremental claim amounts are statistically independent."
Does it mean the same thing as "future development of a cohort is independent of historical factors" which is Mack model assumption.

 

[Ian Senator]

They certainly sound very similar or the same, don't they. Although I think how those assumptions are implemented into the model are slightly different.

Mack is analytical, closely based on the BCL method. It assumes that the sequential development factors are not correlated.

Bootstrapping is simulation. Don't forget bootstrapping is a process rather than a unique method. So you can even bootstrap the Mack method (best not go there!). When you're sampling for your simulation, you can make assumptions when sampling such as independence. That's why the assumption looks similar for the bootstrap that's referred to in the Core Reading.

If you're really interested in the finer detail, I suggest you look at Hindley's book (referred to by the Core Reading). But you need your stats head on - it's far more detailed than you need for the exam. Thankfully past exam questions on these assumptions have required little more than verbatim cut and paste Core Reading, so perhaps we're worrying about nothing.

Hope that helps
Ian