Hi there, I have a question about the QA Bank 3.11 part c) Poor claims experience throughout 2013. The solutions say that there is no distortion in the pattern , which cannot be true. The triangle is on an AY basis, so poor experience (assuming to mean high claims volumes = high claims severity) on 2013 will only affect those claims that occur in the 2013 AY. If these distorted claim figures are used within a simple CL then they would of course distort the pattern. I've attached the Excel 2013 example so I don't muck the explanation up.... Unless I have got the definition of "poor experience" wrong, I can't see how this wouldn't affect the pattern... Cheers, Alun.
Hi there The solution says that it wouldn't cause a distortion in the effectiveness of the chain ladder not that the pattern itself wouldn't be distorted. The question says poor claims experience throughout 2013. This is saying that the same pattern is still expected from development period to development period, although the absolute level of the claims at each development period is higher (eg 10 times say). If every point is ten times higher than normal then the individual link ratios will still be the same. As your example correctly shows, if you use a column sum average approach, then the overall pattern maybe distorted as you will give more weight to the pattern from 2013. Note that if all years develop exactly the same (which is one of the underlying assumptions of the BCL) then this increased volume of claims in 2013 wouldn't actually change the overall pattern. It only has an effect because the patterns vary by year and you are effectively giving different weights to each year. However, the question says that the company uses the IACL method, so you need to be careful about talking about using column sum averages in this specific case!
