M. Berens, consider the insightful points you made well taken. I agree that
it will be unlikely to find everything one system. In fact, I believe the
last company I worked for had three main systems, 1) Accounting, 2) Policy
Registration and Issuance, and 3) Investment tracking. The
registration/issuance system automatically generated files to be sent to ISO
for statistical reporting and also generated a bunch of fairly crude, paper,
"management" reports. All the "good" info flowed from the actuarial
department, including pricing, reserving, and profitability. We used SQL
based programs to query mainframe databases combined with reports form
accounting and investments.
The total system worked well, but had some problems. For instance the product
managers relied on the generated reports to analyze their books. And many
times their unit was profitable on a calendar+case basis but not on an
accident year +case +ibnr basis (a report which came from actuarial).
Unfortunately, bonuses where based on the latter. Nothing can cause friction
like reduced pay! Also there were classic system errors like a database field
labeled "net written premium" which was actually direct or the database that
contained direct, ceded, and assumed in one field (premium), differentiated by
a different and misleadingly named field (invariably first time users
overestimated premium).
Our current system is pretty rough to use. We rely heavily on system
generated printed reports (data entry ugh!), also multiple systems are in play
and the knowledge base on these systems is shrinking. Very few people know
how (or why) the system works. Plus the interface is archaic, old mainframe-
type screens and no documentation exists. The commands and other information
are passed on as part of an oral tradition! : ) Yeah, funny until you have to
try and assemble data from all that!
On a lighter note, I am thinking that the CAS should have an exam 13 on data
integrity. It would be like the (star trek) starship officer cadet training
where the captain-to-be faces a disaster. The CAS equivalent could be
handing a "rock solid" report to the CEO (after days of datamining & double-
checking), who then declares "clearly this number is too big, go figure what's
wrong" Of course CEO doesn't care that a field was misleadingly named.
Take care,
-dpowell
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.