Actuarial Outpost Reserving question - Developing basic layers
 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

#1
06-07-2012, 03:11 PM
 formose Member CAS Join Date: Jun 2012 Studying for yolo Posts: 878
Reserving question - Developing basic layers

(Bear with me for a second as I fumble through asking this, I'm only an intern)

In calculating reserves I'm separating claims into a basic layer (up to the first 300k of a claim) and an excess layer (everything over 300k).

Note that because of this construction some basic claims are "maxed" - they are already over 300k so they cannot develop any more.

These claims are of course present in the triangles and contribute to the development factors.

My question is, when applying development factors, do you first subtract out maxed claims, apply the development factor, then add maxed claims back in,

OR

simply apply the development factor to the full amount including the maxed claims.

I am leaning towards the second since although maxed claims cannot develop they compensate for this by bringing the development factors down appropriately, but the actuary I am working under uses the first (he told me to ask here since I couldn't find any literature regarding this sort of split of the data).

Last edited by formose; 06-08-2012 at 10:55 AM..
#2
06-07-2012, 03:38 PM
 MountainHawk Member CAS AAA Join Date: Dec 2001 Location: Salem, MA Studying for Nothing!!!! College: Lehigh University Alum Favorite beer: Yuengling Posts: 64,850

I am under the opinion that if the entire triangle is of basic claims only, you are ok to use the development factors directly.

However, if there is a significant trend over the relevant development period, you may wish to detrend your limits and not just use the raw data to develop factors.

For example, if there is a 5% trend, for AY 2012 your limit is 300K, for AY 2011 make it 286K (300/1.05) for 2010 make it \$272K, etc.
__________________

Play in the AO Prediction Game now!

1
#3
06-07-2012, 03:40 PM
 Buck Member CAS AAA Join Date: Mar 2005 Posts: 3,773

Since this is your very first post on the forum, you must have a legit reserve question. However, "bare with me" means what? Are you an attractive young female person? Otherwise, I'd stay away from your invitation.
#4
06-07-2012, 04:16 PM
 Dark Horse Member SOA AAA Join Date: Jun 2012 Location: SoCal Posts: 129

Quote:
 Originally Posted by Buck Since this is your very first post on the forum, you must have a legit reserve question. However, "bare with me" means what? Are you an attractive young female person? Otherwise, I'd stay away from your invitation.

Even if it were phrased "bear with me...", he/she could be suggesting the type of activity that is going on in ShakeNBakes Avatar. I don't know if I'd accept that invitation either.
#5
06-08-2012, 01:49 AM
 Buck Member CAS AAA Join Date: Mar 2005 Posts: 3,773

Now, bear with me for a sec ...

For anything, you must be familiar with the situation first ... however, given the presumption that what you described above is accurate, you should cap (or "maxed" in your language) all individual claims at the limit, and separate the original triangle into two: capped & excess. Then, develop the capped triangle as you regularly do, and finally add the excess back without development. The large claims are usually due to special situations and they do not develop. If you keep them in without adjustments, they will make your link ratios wacky (there is no "self-adjustments" as you think). Those large claims were regular claims until lawsuits turn them into million\$ ones. They have giant developments from usual to unusual then no development. If you keep them mixed in, their dollar amounts will drive the link ratios wrong.

So the odds are the actuaries you work with are right and you're wrong.
#6
06-08-2012, 10:54 AM
 formose Member CAS Join Date: Jun 2012 Studying for yolo Posts: 878

Ok Buck and MountainHawk you both seem to agree on that you should just develop the basic triangle regularly.

But @Buck - do you actually do that with excess claims? Just don't develop them? Sure, the excess portion of claims usually is due to litigation but surely you should develop them in some manner since they will probably continue to incur expense until they are eventually closed.
#7
06-08-2012, 11:05 AM
 MightySchoop Member CAS AAA Join Date: Aug 2009 Location: Insula Primalis Posts: 7,287

Quote:
 Originally Posted by formose Ok Buck and MountainHawk you both seem to agree on that you should just develop the basic triangle regularly. But @Buck - do you actually do that with excess claims? Just don't develop them? Sure, the excess portion of claims usually is due to litigation but surely you should develop them in some manner since they will probably continue to incur expense until they are eventually closed.
Without knowing what line you're working in, and what the practices of your claims department are, it is pretty difficult to give you a recommendation of what to do on those excess claims.
__________________
Dare to believe you can survive
#8
06-08-2012, 11:32 AM
 formose Member CAS Join Date: Jun 2012 Studying for yolo Posts: 878

Quote:
 Originally Posted by MightySchoop Without knowing what line you're working in, and what the practices of your claims department are, it is pretty difficult to give you a recommendation of what to do on those excess claims.
It's all homeowners' stuff. I'm not so interested in the technique for applying it since the past reserving has never used that technique I'm just interested from a purely educational standpoint since it was never mentioned in Friedland (the only reserving text I've read).
#9
06-08-2012, 04:07 PM
 whisper Member CAS AAA Join Date: Jan 2002 Location: Chicago Favorite beer: Hefewizen Posts: 38,897

Depends on what you've got.
If you're using limited LDFs, then just apply the factors.

If you're estimating limited LDFs, then I'd do it both ways.
+ Apply the factors
+ Remove the limited losses, apply the factors, then add them back in
The answer will probably lie between the two.
#10
06-11-2012, 06:08 PM
 jesusislord Member CAS AAA Join Date: Aug 2006 Posts: 2,787

i generally look at what each method gives me and determine if an adjustment is appropriate, ie if the paid method is way higher than the incurred due to a large payment on a claim in an immature period.

i also look at the ibnr relative to open claims and case outstanding. but i mainly work with work comp and gl stuff.
__________________
God is real and His love is unconditional.