mages / ChainLadder

Claims reserving models in R
https://mages.github.io/ChainLadder/
77 stars 63 forks source link

Mack se recursion rewrite #53

Open chiefmurph opened 6 years ago

chiefmurph commented 6 years ago

To address my new issue 15, I rewrote the MackRecursive.S.E to be row-wise r.t. column-wise. Now when a development age (column) has more than one row in the triangle whose latest value is at that age, the Mack statistics calculate correctly -- the values for the last two rows below are identical and the Mack.S.E value for the second row is no longer zero:

G <- rbind(GenIns, 11=GenIns[10,]) MackChainLadder(G) MackChainLadder(Triangle = G)

  Latest Dev.To.Date  Ultimate      IBNR  Mack.S.E CV(IBNR)

1 3,901,463 1.0000 3,901,463 0 0 NaN 2 5,339,085 0.9826 5,433,719 94,634 71,835 0.759 3 4,909,315 0.9127 5,378,826 469,511 119,474 0.254 4 4,588,268 0.8661 5,297,906 709,638 131,573 0.185 5 3,873,311 0.7973 4,858,200 984,889 260,530 0.265 6 3,691,712 0.7223 5,111,171 1,419,459 410,407 0.289 7 3,483,130 0.6153 5,660,771 2,177,641 557,796 0.256 8 2,864,498 0.4222 6,784,799 3,920,301 874,882 0.223 9 1,363,294 0.2416 5,642,266 4,278,972 970,960 0.227 10 344,014 0.0692 4,969,825 4,625,811 1,362,981 0.295 11 344,014 0.0692 4,969,825 4,625,811 1,362,981 0.295