Actuarial Outpost > SoA Module Exercise - Reserve Part
 Register Blogs Wiki FAQ Calendar Search Today's Posts Mark Forums Read
 FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions

 DW Simpson International Actuarial Jobs Canada  Asia  Australia  Bermuda  Latin America  Europe

#11
06-16-2008, 09:07 AM
 DudeMan Member SOA Join Date: Jan 2006 Location: teh Po' Studying for shts&ggles Favorite beer: Ginger Beer Posts: 11,612

Quote:
 Originally Posted by Shadowcat did anyone else get the same as streetballer? i'm actually getting BF in the middle. (IC lowest). I also looked at historical trends and am thinking of using something lower than 12%.
I thought by definition BF is between Incurred and CF methods. Either way, my Incurred method has lowest, followed closely by BF and then CF is by far the highest. I'm not sure how reliable the CF method is for this project. Claims lag is 6 months. Only 1 out of 6 months has CF greater than 80% and 1 other > 70%. Normally, I would be looking to set a pure premium for 4 of the 6 months.
__________________
What difference, at this point, does it make?
#12
06-17-2008, 11:08 PM
 Golden Monkey Join Date: Aug 2004 Posts: 21

This could be a dummry question.

I try to understand better in the case study Incurred Claim method, when it trends in 12% annually, why starting from July, the formula is = (1+12%)^(3.5/12)?

Where this 3.5 comes from?

Confusing!!!
__________________
Our greatest glory is not in never falling, but in rising every time we fall. -- Confucius
#13
06-18-2008, 08:20 AM
 DudeMan Member SOA Join Date: Jan 2006 Location: teh Po' Studying for shts&ggles Favorite beer: Ginger Beer Posts: 11,612

Quote:
 Originally Posted by Golden Monkey This could be a dummry question. I try to understand better in the case study Incurred Claim method, when it trends in 12% annually, why starting from July, the formula is = (1+12%)^(3.5/12)? Where this 3.5 comes from? Confusing!!!
They use a 6 month average from 1/1/2005 - 6/30/2006. Midpoint = 4/1/2005. You are trending from 4/1/2005 to the midpoint of July '05 (7/15/2005) for a total of 3.5 months.

I did mine using 12 months rather than 6. I also used 12.9% (CY 2005 / CY 2004) trend rather than 12%
__________________
What difference, at this point, does it make?
#14
06-18-2008, 11:38 AM
 gppbb Member Join Date: Sep 2005 Posts: 245

I am using 200401 thru 200512. How come I see a negative trend from 2004 to 2005? Did anyone else see the same? I think for conservative purpose, I will use a positive trend, but I have a hard time justifying a 10% trend given a negative trending observation.
#15
06-18-2008, 12:31 PM
 DudeMan Member SOA Join Date: Jan 2006 Location: teh Po' Studying for shts&ggles Favorite beer: Ginger Beer Posts: 11,612

Quote:
 Originally Posted by gppbb I am using 200401 thru 200512. How come I see a negative trend from 2004 to 2005?
Bad idea. Only half of the months in 2005 have completed claims. Closest you should use is 200406-200505.

Quote:
 Originally Posted by gppbb Did anyone else see the same? I think for conservative purpose, I will use a positive trend, but I have a hard time justifying a 10% trend given a negative trending observation.
Some months do have negative trend from 2004 to 2005. Coincidentally, those also happen to be the months that had huge positive trend from 2003 to 2004.
__________________
What difference, at this point, does it make?
#16
06-18-2008, 04:24 PM
 gppbb Member Join Date: Sep 2005 Posts: 245

Quote:
 Originally Posted by GraffixMan Bad idea. Only half of the months in 2005 have completed claims. Closest you should use is 200406-200505. Some months do have negative trend from 2004 to 2005. Coincidentally, those also happen to be the months that had huge positive trend from 2003 to 2004.

Guess I wasn't clear on what I meant. I used 200401-200506 for developing the CF stuff. The second half of 2005 was just used to calc paid claims.
#17
01-16-2009, 01:57 PM
 BurnTheMidnightOil Member SOA Join Date: Dec 2004 Posts: 123

Did anyone else notice a funny error in the model solution spreadsheet for Task1 - CFM method?
#18
02-17-2009, 09:41 PM
 DaBears Member SOA AAA Join Date: Oct 2006 Location: In my happy place Posts: 182

Quote:
 Originally Posted by BurnTheMidnightOil Did anyone else notice a funny error in the model solution spreadsheet for Task1 - CFM method?
Yes
#19
03-06-2009, 12:05 PM
 Kram Member Join Date: Sep 2006 Posts: 76

hmmm... i'm having the same trend problem that gppbb had
i suppose i'm the only one working on this thing now ... anybody out there?
#20
05-31-2009, 11:49 AM
 stallion78 Member SOA Join Date: Oct 2007 Location: Buffalo Wild Wings Studying for FSA Modules Favorite beer: Hobgoblin or Stella Posts: 46

I agree with everyone on the trend issue. Due to lack of details in the EOM Exercise Overview, we have no basis for determining what could be causing such a strange trend pattern (a sudden increase in level of managed care could be bringing down trends, change in mix of services, favorable provider contracting, etc), and therefore, have little basis to go off of when determining what an appropriate trend value is when calculating IBNR.

Has anybody looked at the difference between Medical and Rx claims? The trend pattern is dramatically different. Also, the Rx lag pattern doesn't even come close to matching reality. Medical claims is showing a fairly similar lag pattern to that of Rx claims. Rx claims should be at least 90% complete after the second month of runout. I wish the SOA would spend a little more time making sure their exercises were reasonable reflections of reality.