MultiBootMackChainLadder: MultiBootMackChainLadder

Description Usage Arguments Details Value See Also Examples

Description

The multi boot mack chain ladder algorythme computes todays and next-year common estimates on a portefolio of several triangles, following closely a synchronised version of BootMackChainLadder.

Usage

1
2
MultiBootMackChainLadder(triangles, B = 100, distNy = "normal",
  names = NULL, seuil = NA, stab = NA)

Arguments

triangles

A List of Triangles objects of the same size.

B

The number of boostrap replicates

distNy

The process distribution for next year increments. Either "normal" (default), "residuals.bycolumn","residuals.global" or "residuals". See details below.

names

enventual names of the different triangles. IF set to NULL, the procedure will try to get names from the triangles list (first argument)

seuil

Eventual exclusions limit for residuals. Set to NA (default) to avoid excluding anything.

Details

This function calculates, on a list of triangles, a synchronised bootstrap in the Mack model for the triangles. The One-year point of view is also calculated. It returns a specific S3 object, on wich you can use the functionsmean, CDR, NyIBNR, Corel provided by the package.

This model uses the fact that the Mack model can be seen as a quasi-glm, and therefore provide resiuals. Bootstrapping thoose residuals on the upper-left triangle allows to get bootstrap distribution of today's estimations (reserves, ultimates, ...). Furthermore, if you simulate net year payments with a given process ditribution in each simulation, it gives 1 year results.

If distNy = "normal", it follows *boumezoued & al* and converges to the Merz-W<c3><bc>thrich formula in the Braun model. If distNy = "residuals", it also converges strongly but NOT to the Merz-W<c3><bb>thrich formula. Same for ohter possibilities. Notes that if you choose distNy = "residuals.bycolumn", the residuals will be resampled inside each column and not accross columns. Setting distNy = "residuals.global" or "residuals" triggers the same code.

Value

a MBMCL object containing a list of BMCL objects and a little more.

See Also

BootMackChainLadder

Examples

1
2
3
data(ABC)
triangles <- list(tri1 = ABC, tri2 = ABC, tri3 = ABC)
MultiBootmackChainLadder(triangles,100)

lrnv/mbmcl documentation built on May 24, 2019, 2:52 p.m.