How do we percentile match with truncation? For example, say we are given loss amounts of 300 350 450 500 600 and told our observations were truncated at 250 and asked to percentile match to exponential using 50%(making this up on the fly)...
Would we use 450 as our 50th percentile? Or somehow incorporate the truncation? (perhaps our 50th percentile = F(450)/S(250))
Would we use 450 as our 50th percentile? Or somehow incorporate the truncation? (perhaps our 50th percentile = F(450)/S(250))
Percentile Matching with truncation
