How to create a compound item in the result template?

Hello,

I am interested in modelling a compound item such as :

(Gross Loss + Reinsurance result)

and be able to find out different level of VaR or TVaR.

How should I proceed?

Kind regards

Jean-Michel Piat

Jean-Michel,

imagine you run a simple model with one claims generator and one reinsurance contract only. I selected an attritional claims generator with a exp(1) distribution and a 50% quota share.

In the result view, I can directly observe tvar (etc.) of

• claim lean development gross of tvar 4.69
• claim lean development net tvar 2.34 (i.e. after quota share)
• quota share tvar 2.34

and many more. If I choose the claims to be logn(1, 10) with the same reinsurance, I can observe a similar behaivour but different numbers. As expected for a quota share.

Does this answer your question or did I missunderstand your question? 

For direct contact, please find my contact details here: Re: Reserve Distribution

Regards,

Markus M

Jean-Michel,

Let me give you another example: I've built a small parametrization with two claims generators only. Lets have a look at their mean and 99.5-TVaR.

As you will see, the TVaR for the sum of the claims generators (1) is smaller than the sum of the TVar of both claims generators (2) and (3). The effect of diversification.

Now, you might want to see other groups than those provided per default, i.e. other views of the diversification effect. As a starting point, you can change the Default view (4) to what is better suited. If you still don't see your desired view, you could do many things of which I suggest two alternatives:

• Build your own structure in the parametrization (no screen-shot here, but this would include the elements you are analyzing), allowing you to filter (5) afterwards for what you are looking for; or to 
  
• write your own view. This, however, takes some coding, please check out this link for further details:
  https://issuetracking.intuitive-collaboration.com/jira/browse/PMO-1033
  

Hoping this helps

Regards,

Markus