ASM 8th 15.2

[langstafftigerpizza]

Question goes like the following:
The number of claims on an insurance coverage has a negative binomial distritubion with mean 2 and variance 6. The claim size distribution is binomial with parameters m = 3, and q = 0.4. A reinsurance contracts pays aggregate claims over an aggregate deductible of 2. Determine the expected aggregate loss paid by reinsurance.

My question is: in solution, it modifies the claim size distribution (negative binomial) to exclude claim size of 0. Can someone explain the reason behind that? and also why modify only claim size distribution not the frequency distribution?

Thanks

[silvergrey]

Question goes like the following:
The number of claims on an insurance coverage has a negative binomial distritubion with mean 2 and variance 6. The claim size distribution is binomial with parameters m = 3, and q = 0.4. A reinsurance contracts pays aggregate claims over an aggregate deductible of 2. Determine the expected aggregate loss paid by reinsurance.

My question is: in solution, it modifies the claim size distribution (negative binomial) to exclude claim size of 0. Can someone explain the reason behind that? and also why modify only claim size distribution not the frequency distribution?

Thanks

I have the same confusion here. Need somebody help.

[Abraham Weishaus]

In all "aggregate payments" questions, there are two ways to solve the problem:

(1) Calculate average size of losses times average number of losses. (The first average would have to include losses of zero.)
(2) Calculate average size of payments (i.e. losses>0) times average number of payments.

I chose to do it the second way, so there is no need to modify the loss size distribution.

[Actuarialsuck]

You know you can delete posts right?

[langstafftigerpizza]

In all "aggregate payments" questions, there are two ways to solve the problem:

(1) Calculate average size of losses times average number of losses. (The first average would have to include losses of zero.)
(2) Calculate average size of payments (i.e. losses>0) times average number of payments.

I chose to do it the second way, so there is no need to modify the loss size distribution.

thanks. :)