=?ISO-8859-1?Q?Re:=20B=FClhmann=5Fapproach?=

VanGulik@aol.com
Sun, 4 Apr 1999 19:17:46 EDT

In a message dated 4/2/99 3:32:57 PM Eastern Daylight Time,
droidakis@MORNEAUSOBECO.COM writes:

<< I am wondering how to calculate the VHM because in some case VHM=TV - EVPV
and in other case they calculate the VHM as Summation [prob. of Xi * (Xi -
E(x)) ]. Why sometimes it's the first formula that they use and sometimes
the second one ?
>>

If you have actual data, you can't calculate the VHM directly. here's some
data
in 2 column format:

number of claims, count of policies having *x* number of claims

0 claims, 700
1 claim, 200
2 claims 60
3 claims, 30
4 claims, 20

assume the population has two types of risks, bad risks, and good risks. The
VHM cant be determined because there is no information on the hypothetical
means. for example a good risk could have o claims and a good risk could
have 4 claims. The number of claims is not the category, Good risk/Bad risk
are the categories.. See 5/85 q38 for an example.

Somebody feel free to chime in. I'm a bit shakey on this.

VG