View Full Version : Part 5: Marker/Mohl

12-03-2004, 02:12 PM
I'm reading this article and I came across the portion of the paper describing how to account for exposure on a claims-made basis. For example, on page 281, Example 1 says that one exposure is contributed to each lag of year j. Does this make sense? Given that a policy is "mature" after four years, should we assume that an insurer is equally exposed to claims reported this year occuring 3 years ago as it is to claims reported this year occurring this year? Maybe it's discussed further in the paper, but it sounded like the exposure discussion was fairly wrapped up. Thanks for your thoughts.

Wigmeister General
12-03-2004, 02:19 PM
Disclaimer: It's been many years since I've read that paper.

I believe the idea was that there's a mismatch between the exposure base and the actual exposure. The concept is that claims reported in a particular year (for mature policies) should be similar to claims reported in a previous year or subsequent year (but for claim cost trend).

The premium charged on old occurrence policies could not properly predict all of the major changes in future settlement patterns, but the premium charged for mature claims made policies was less susceptible to changes in long-term trends.

12-03-2004, 02:30 PM
Your post makes sense, but it concerns the general concept of the paper. What isn't lining up to me is what seems to be an oversimplifying assumption.

I'm an insurance company rating policies on a claims-made basis, where maturity is established after 4 years. Therefore, I am exposed to 4 different groups of claims: those occurring this year, those occurring last year, those occurring 2 years ago, and those occurring three years ago, all of which are reported this year. Given a claim is reported this year, is it equally likely that it is from any of these four groups? It seems like you're assuming too much, by answering yes.

Howard Mahler
12-03-2004, 02:33 PM
They will be estimating pure premiums for each report year lag combination. We do not expect these pure premiums for different combinations to be the same.

One exposure at lag 1 is not expected to produce the same frequency or pure premium at lag 2. An exposure at one lag is a whole different thing than an exposure at anothe lag.
So this differs from our usual use of exposure.

Howard Mahler

P.S. Marker and Mohl should be read after most of the other papers.

12-06-2004, 11:25 AM
P.S. Marker and Mohl should be read after most of the other papers.
I'm not reading anything yet, but I should be starting shortly after Christmas. Where should I begin?

12-06-2004, 11:38 AM
I think it's good to start with the McClenahan Ratemaking paper. It's a good overview.

Wigmeister General
12-06-2004, 11:40 AM
Followed by the Finger paper (if it is still on the syllabus).

12-06-2004, 01:57 PM
I'm going in the order of the 2004 All 10 manual. Should there be any material differences between the 2004 and 2005 versions?

Howard Mahler
12-06-2004, 02:04 PM
I think it's good to start with the McClenahan Ratemaking paper. It's a good overview.

I agree start with McClenahan!
It is about 13% of the exam and basic stuff you have to know.
It will help you understand the other papers.

Also leave Tiller/Sherwood to near the end of your first pass.