Premium Trend Periods
Glen_R._Calihan@westfield-cos.com
Mon, 2 Mar 1998 12:52:08 -0500
I concur with the two respondents to the Sweeney question.
As long as we are moving from earned/incurred data to a future
earned/incurred basis, then the dates and periods should be the same.
Sometimes, though, the wording that some of the authors use, like those
quoted by one of the respondents, can be a little confusing. Like for
example, Feldblums 'the midpoint of the future policy period'. He is right
in his statement but you still have to determine properly what that period
and midpoint is (ie length rates are to be in effect, policy period).
Compare to Homans ' to one year beyond the effective date". The authors
are in agreeement under one circumstance (annual policies, rates effective
for one year)but not under all circumstances.
We had this issue come up in our department a number of years ago. There
were some on each side of the argument/discussion. I am not familiar with
the Walters paper but it may have had some influence on some. Also the way
ISO words some of their descriptions of this issue can be misleading (I am
told) but when you look into it in some detail, their procedures are
consistent with the same period approach. Here, we all now do indications
the same way, one trend period for both.
Glen Calihan