---Yin
----------
From: Mary_Frances_Miller@sedgus.com
To: casnet@lists.casact.org; GaryBalto@aol.com; Lawn;Yin
Subject: Re[4]: [Fwd:CASNET]
Date: Tuesday, June 23, 1998 8:51AM
Sure, you can always calculate a number based on claim counts. But
all those calculations assume that there is a reasonable a priori
choice to take the complement of credibility. Does that assumption
hold in surety? I've never worked with the coverage, so I don't
know,
but if the current rate is the only other information you've got
and
it's based on economic conditions which no longer hold, full
credibility to the current experience may be a better choice.
Mary Frances
______________________________ Reply Separator
_________________________________
Subject: RE: Re[2]: [Fwd:CASNET]
Author: "Lawn;Yin" <Yin.Lawn@cna.com> at _internet
Date: 6/22/98 4:59 PM
I some how disagree. Classical credibility theory can give you an
absolute number as long as you assume that actual and the expected
result will fall within X percentage with the probability less than Y.
I am not aware of any credibility standard for Surety but that doesn't
mean that they can't be derived from same idea that drives the
credibility standard for Auto. Using classical credibility theory, you
can obviously construct a confidence interval to figure out number of
claims required in order to have the actual result to be within say, 10%
of the expected result at least 95% of the time. However, I suspect if
you take this approach you'll get a number that's too high to be
practical. Giving the severity and frequency natural of the Surety
claims, it might be impossible to achieve that kind of confidence and
precision.
Another approach, which is probably better, is to take a Basian
approach. This might be more practical because it gives you an optimal
result. Instead of giving you an absolute credibility, it tells you
how to get the best out of what you got(it might still be bad but at
least it's the best you can get). This is best especially when none of
your calculations seem credible. Basian basically compares the results
of various methods(look at expected value of process variance and
variance of hyperthetical mean) and assign credibility accordingly.
Of course, subjective judgments are very important as well, especially
when credibilities are low. Surety risks are extremely
"individualized", that is risks are highly depended on company's
individual characteristics such as loss control procedure. I would also
check with experienced underwriters to get some of their puts on what is
considered to be fully credible.
---Yin
----------
From: Mary_Frances_Miller@sedgus.com
To: casnet@lists.casact.org; GaryBalto@aol.com
Subject: Re[2]: [Fwd:CASNET]
Date: Monday, June 22, 1998 9:32AM
Does the SAA believe its numbers are fully credible because there is
nothing
appropriate to weight against? Credibility, after all, just gives you
relative
weights and isn't an absolute number.
Mary Frances Miller
______________________________ Reply Separator
_________________________________
Subject: Re: [Fwd:CASNET]
Author: <GaryBalto@aol.com> at _internet
Date: 6/19/98 6:05 AM
As you already know the SAA believes their numbers are fully credible
for each
business segment they use to generate a loss cost. I don't know that
you will
find a "standard" in the literature
In a message dated 98-06-18 12:16:45 EDT, you write:
<< Paul,
I need to post a question on CASNET. Does anyone know what type of
credibility
standard in used in Surety? How is partial credibility determined?
I've
already checked with the SAA but to no avail.
Thanks.
Marlene >>
Visit the CAS Web Site at http://www.casact.org
===============================================
To subscribe or unsubscribe from CASNET:
Send an e-mail to caslists@lists.casact.org
Type in the body join casnet to subscribe
or leave casnet to unsubscribe.
Visit the CAS Web Site at http://www.casact.org
===============================================
To subscribe or unsubscribe from CASNET:
Send an e-mail to caslists@lists.casact.org
Type in the body join casnet to subscribe
or leave casnet to unsubscribe.
Visit the CAS Web Site at http://www.casact.org
===============================================
To subscribe or unsubscribe from CASNET:
Send an e-mail to caslists@lists.casact.org
Type in the body join casnet to subscribe
or leave casnet to unsubscribe.
Visit the CAS Web Site at http://www.casact.org
===============================================
To subscribe or unsubscribe from CASNET:
Send an e-mail to caslists@lists.casact.org
Type in the body join casnet to subscribe
or leave casnet to unsubscribe.
Visit the CAS Web Site at http://www.casact.org
===============================================
To subscribe or unsubscribe from CASNET:
Send an e-mail to caslists@lists.casact.org
Type in the body join casnet to subscribe
or leave casnet to unsubscribe.