To answer Michael's other question, methods 2, 3, & 4 all assume that
the future development will be the same for all accident years. If you
were to break out the expected development by period, all the accident
years would show the same expected development by period. The
difference in the expected ultimate losses for each AY comes from the
known development of each AY. So, unless the projected ultimate result
was modified for inflation for each AY, the methods will under-predict
the true ultimate losses. Stanard does make this adjustment in
Exhibits IV & V.
On Thu, 15 Oct 1998 08:33:17 -0400 "Sce, Michael"
<MAS2@ChartwellRe.com> wrote:
> Jason's answer makes sense except that the reviewer, John Robertson, says on
> page 150 that Stanard does include inflation from report to payment.
>
> Stanard says on page 133 that "...we expect methods 2, 3, & 4 will
> under-predict, because they all implicitly assume that expected losses by AY
> are the same, which, with inflation, is not true."
>
> It's not clear to me what Stanard means by that statement. Does he mean
> even with inflation taken into accout it's not true or because of inflation
> it's not true?
>
> Anybody?
>
> Second queston involves the upward bias in the LDFs. That's the best I can
> do there. I am not aware that Stanard offers a reason why. In method #1,
> he talks about how the prediction error is skewed and, presumably, high.
> The other methods also follow similar patterns, probably with no skewness
> and smaller prediction errors.
>
>
>
>
>
>
> > -----Original Message-----
> > From: Jason T. Sash [SMTP:Jason.T.Sash@EMCIns.Com]
> > Sent: Friday, September 04, 1998 8:04 AM
> > To: Jim Shoenfelt
> > Cc: Exam7
> > Subject: Re: Stanard questions
> >
> > The reserving portion of Stanard's model only accounts for inflation
> > at the time of reporting and does not make any adjustment for the
> > inflation between the time of reporting and payment. Essentially it
> > reserves claims at the amount that they would be settled for if they
> > were paid when they were reported and not at the later actual payment
> > date.
> >
> > I'm not completely sure about the second question. I believe that it
> > might have to do with the same bias in method 1, the age-to-ultimate
> > factors (f). Since they produce an upward bias for method 1, I can
> > only assume that they also produce an upward bias for the Modified B-F
> > method as well. Someone else can either confirm this or provide
> > another answer.
> >
> >
> > On Thu, 03 Sep 1998 21:22:46 -0400 Jim Shoenfelt
> > <jimshoenfelt@amdyne.net> wrote:
> >
> > > Why does Stanard's method of generating losseson oaverage lead to
> > underreserving by (Ip/Ir) - inflation betweenreport & pmt?
> > >
> > > And why does Stanard say the the BornhuetterFerguson modified method has
> > upward inherent bias in each age-to-age ultimate prediction?
> > >
> >
> > ----------------------
> > Jason T. Sash
> > Jason.T.Sash@EMCIns.com
----------------------
Jason T. Sash
Jason.T.Sash@EMCIns.com