Choosing the two trend values wisely can
produce credible results. By letting the
system calculate the two trend factors
on data from August 2008 through December 2010, and comparing that with
actual results, I get an R-squared value
of .927—a very high positive correlation.
In tracking results without the benefit of
knowing what the best trend values are, I
produce results with an R-squared value
of .884 for the same time period.
Figure 2 shows the relative accuracy
between picking the trends at the time
of calculation versus those trends that
represent what could have been, with
hindsight, 100 percent accurate in picking
those values. ( I chose that period because
it represented the most stable time of my
disbursement system issues.)
The speed-up of payments in which my
company is engaged ultimately will make
the Month 1 completion factors more
stable. Still, the Month 1 completion fac-
tors always will be more volatile than
Month 2 (and Month 2 more volatile
than Month 3, and so on). Both the pay
cycle method described by Andrew Sa-
vona and the pertinent day process that
I describe seek to enhance the predict-
ability of the ultimate claims incurred for
a particular month. Although my data are
irregular because of the increase in pay-
ment speed embedded within, I believe
both techniques can work together to
produce increased predictability.
100 percent accurate result. It would be
interesting to see the optimum weighting
of the two methods necessary to produce
the most accurate results. I will work on
doing that once I have enough data points
sufficiently timed not to be influenced by
our disbursement system issue.
I enjoy the challenge of creating accurate claims reserving estimates. The
value of increased accuracy in claims
reserving not only makes for more accurate financial statements; it produces
more accurate completion factors, restates, and PPs. These gained accuracies
benefit not just the financial statements
but also financial reporting, forecasting,
rate-making, and other miscellaneous reporting activities.
LOUIS KORB is director of corporate
actuarial for Medical Mutual of Ohio
This article is solely the opinion of its author.
It does not express the official policy of the
American Academy of Actuaries; nor does it
necessarily reflect the opinions of the Academy’s
individual officers, members, or staff.