CH7 practice question 7.1(iii) (a)

[kiki]

Hi,

I am having issue of understanding the answer for Q7.1(iii) part (a) about different ways of defining large individual claims, can someone please help me ?

Method 2: "extract the whole of each large claim and associate history if its incurred claim amount exceeds a certain threshold..."

• so am i right to say it is looking at the current incurred amount and flag the claim as large if incurred is greater than large loss threshold ?
• in the point of " difficult to reconcile with last year's data" : is that because non-large triangle will need to be restated after filtered the large claim, therefore the triangle of non-large from previous years are difficult to reconcile with current year ?

Method 3: "once large always large"

• This method is different to method 2 as it is looking at the historical incurred instead of current view of incurred ?
• if it is right , then how do you allow the inflation effect? as claim incurred as 500k 10 years ago may classified as large , but now the large threshold may need to increase to 1 million?
• is that why the last point "distort any large claim average cost analysis ? as number of claims can be classified as large will increase year over year due to inflation

Method 4

• I am not quite understand the last point about "development factors that rely too heavily on such an instance would result in an optimistic non-large IBNR estimate ? so is that mean using such method may lead to under-estimate the IBNR for non large claim? how so?

thank you so much for your time and patient

kiki

 

[Busy_Bee4422]

Good day

Behind this question is the basic issue of dividing your claims into homogenous groups (large and attritional) before triangulation, the specifics of how you would go about it and how you would assess the quality of the grouping and its impact on the results.

Method 2

• Yes
• Yes. Before the claim was classified as large it was in the attritional claims. The attritional claims triangle will need to be restated.

Method 3

• I may have missed it but it seems to me the emphasis is not on the threshold of the amount for large claims but in that a claim can be assessed to be large when you take into consideration the amount paid and the amount outstanding. When the outstanding amount has more certainty it may well prove to be non-large. This will result in the movement of claims between the two categories in both directions.
• See the point above. The large triangle will end up with both large and non-large claims. If inflation applies differently to large and attritional claims it will be misstated. Additionally, this will depend on the actual method you employ in triangulation.
• The large claims will now have claims that were thought to become large but actually proved to be non-large. This will reduce your average large claims amount.

Method 4

• This method differs from method 2 in that you don't adjust the associated history. When you move a claim from attritional group to large it happens in the development period in which it was reclassified whilst leaving the history the same. This means that in all prior development periods, this claim is in the attritional claims triangle and then from this period it moves to the large claims triangle. Because the prior period attrition claims total has the history of the now large claim but in this period the claim has been reclassified it will result in a drop in the current period claims total in the attritional claims triangle caused by the reclassification. If this is a significant issue then there may be a significant drop in the development factor influenced strongly by this reclassification rather than paying out of claims. Claims may appear to be developing faster than they actually are leading to underestimating of the IBNR.

 

[Ian Senator]

Thanks to Busy_Bee. Just to clarify method 3 - yes, it's looking at the historical incurred if it was ever above the threshold.