Actuarial Outpost
 
Go Back   Actuarial Outpost > Actuarial Discussion Forum > Property - Casualty / General Insurance
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions


Upload your resume securely at https://www.dwsimpson.com
to be contacted when our jobs meet your skills and objectives.


Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 04-30-2014, 02:55 AM
hacktuary85 hacktuary85 is offline
Member
 
Join Date: Apr 2011
Location: London
Favorite beer: Peroni
Posts: 75
Default Bootstrapping Age-to Age factors

Hello,

Recently I came across an english actuary who was creating a range estimation for reserving by runing simulations resapling the Age to Age factors.

I have never seen something like this before and I cant find any literature about it. All the bootstrap simulations I have ever done were based on resampling errors (mainly pearson errors with a ODP bootstrap).

Did any of you ever done anythng similar to this Age-to Age factor resampling? does it work well? Did he invented it or it is a standard technique?

Thank you
Reply With Quote
  #2  
Old 04-30-2014, 06:59 AM
Ron Weasley's Avatar
Ron Weasley Ron Weasley is offline
Member
CAS AAA
 
Join Date: Oct 2001
Studying for naught.
Favorite beer: Butterbeer
Posts: 8,623
Default

Have you done a literature search on casact.org yet?

I personally have not heard of applying bootstrapping to link factors.
__________________
“Fantasy is escapist, and that is its glory. If a soldier is imprisioned by the enemy, don't we consider it his duty to escape?. . .If we value the freedom of mind and soul, if we're partisans of liberty, then it's our plain duty to escape, and to take as many people with us as we can!”
― J.R.R. Tolkien
Reply With Quote
  #3  
Old 04-30-2014, 08:13 AM
MountainHawk's Avatar
MountainHawk MountainHawk is offline
Member
CAS AAA
 
Join Date: Dec 2001
Location: Salem, MA
Studying for Nothing!!!!
College: Lehigh University Alum
Favorite beer: Yuengling
Posts: 64,850
Default

Unless you have a massive triangle/parallelogram (50 years of history), this seems very limited compared to the traditional England-Verrall style bootstrap.
__________________

Play in the AO Prediction Game now!



1
Reply With Quote
  #4  
Old 04-30-2014, 11:18 AM
hacktuary85 hacktuary85 is offline
Member
 
Join Date: Apr 2011
Location: London
Favorite beer: Peroni
Posts: 75
Default

Quote:
Originally Posted by MountainHawk View Post
Unless you have a massive triangle/parallelogram (50 years of history), this seems very limited compared to the traditional England-Verrall style bootstrap.
In this case almost all the Age-to Age factors would be 1, what is the sense of resampling a set of factors for which 90% are 1's?

Regards
Hack
Reply With Quote
  #5  
Old 04-30-2014, 03:03 PM
MountainHawk's Avatar
MountainHawk MountainHawk is offline
Member
CAS AAA
 
Join Date: Dec 2001
Location: Salem, MA
Studying for Nothing!!!!
College: Lehigh University Alum
Favorite beer: Yuengling
Posts: 64,850
Default

Quote:
Originally Posted by hacktuary85 View Post
In this case almost all the Age-to Age factors would be 1, what is the sense of resampling a set of factors for which 90% are 1's?

Regards
Hack
Not 50 development periods. But 50 diagonals for like 10-12 development periods (so you have 1965-2014 worth of data. Now you can sample from 49 1->2 factors, 48 2->3 factors ... etc.
__________________

Play in the AO Prediction Game now!



1
Reply With Quote
  #6  
Old 04-30-2014, 04:26 PM
jesusislord jesusislord is offline
Member
CAS AAA
 
Join Date: Aug 2006
Posts: 2,787
Default

isnt this there a mclenahan paper out there on this? i dont think its bootstrapping tho.
__________________
God is real and His love is unconditional.
Reply With Quote
  #7  
Old 04-30-2014, 05:37 PM
Guest
Guest
 
Posts: n/a
Default

Quote:
Originally Posted by MountainHawk View Post
Not 50 development periods. But 50 diagonals for like 10-12 development periods (so you have 1965-2014 worth of data. Now you can sample from 49 1->2 factors, 48 2->3 factors ... etc.
Reply With Quote
  #8  
Old 04-30-2014, 05:55 PM
hacktuary85 hacktuary85 is offline
Member
 
Join Date: Apr 2011
Location: London
Favorite beer: Peroni
Posts: 75
Default

Quote:
Originally Posted by jesusislord View Post
isnt this there a mclenahan paper out there on this? i dont think its bootstrapping tho.
Not that I've seen, do you have a link or the name of the paper?
Reply With Quote
  #9  
Old 04-30-2014, 06:16 PM
StephenL StephenL is offline
CAS
 
Join Date: Sep 2010
Posts: 9
Default

I known someone who has published far more complex methods on the topic of reserve variability. He once told me he always starts by looking at the variability of LDFs by development period to come up with an initial feel for the unpaid distribution before progressing on to a more complex process.

Stephen
Reply With Quote
  #10  
Old 05-01-2014, 07:45 AM
MountainHawk's Avatar
MountainHawk MountainHawk is offline
Member
CAS AAA
 
Join Date: Dec 2001
Location: Salem, MA
Studying for Nothing!!!!
College: Lehigh University Alum
Favorite beer: Yuengling
Posts: 64,850
Default

Quote:
Originally Posted by Ke$ha View Post
What, I have 30 years of history for WC, and would have 40 more if some goofball decided that we didn't need the history of outstanding losses when they updated claims systems. It's not that outlandish.
__________________

Play in the AO Prediction Game now!



1
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


All times are GMT -4. The time now is 10:15 PM.


Powered by vBulletin®
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
*PLEASE NOTE: Posts are not checked for accuracy, and do not
represent the views of the Actuarial Outpost or its sponsors.
Page generated in 0.42628 seconds with 9 queries