Hi. I am trying to estimate IBNR for uniformly risk distributed LoBs (say Health, Motor Own Damage etc.,) using both underwriting year and AY origin cohorts. Across LoBs, even after adjusting UW origin cohorts for earning, I end up getting more IBNR than AY origin cohorts. Technically I would expect both to give me similar IBNR, though may not be the same. Has any one else faced the same challenge? If so, how did you circumvent the same? Thanks Sree
An underwriting year triangle will project claim events that have not yet occurred, whereas an AY projection will not. You say you have adjusted your UW cohorts for earning ... an interesting idea but it sounds like your adjustment hasn't particularly helped.