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

Salary Surveys
Property & Casualty, Life, Health & Pension

Health Actuary Jobs
Insurance & Consulting jobs for Students, Associates & Fellows

Actuarial Recruitment
Visit DW Simpson's website for more info.
www.dwsimpson.com/about

Casualty Jobs
Property & Casualty jobs for Students, Associates & Fellows


Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 03-20-2018, 06:20 PM
RockThatScoober RockThatScoober is offline
CAS
 
Join Date: Mar 2018
Posts: 4
Default Include policies with zero paid claims in GLM?

I want to do a frequency/severity analysis for classification using GLM. My data has one entry for each year of each policy, including non-cat claim counts and non-cat paid losses. When doing the severity part, should I exclude policies that have zero paid claims?
Reply With Quote
  #2  
Old 03-21-2018, 11:47 AM
CuriousGeorge CuriousGeorge is offline
Member
CAS SOA
 
Join Date: Dec 2005
Posts: 1,160
Default

Depends on what you did in the frequency part.
Reply With Quote
  #3  
Old 03-21-2018, 12:21 PM
Vorian Atreides's Avatar
Vorian Atreides Vorian Atreides is online now
Wiki/Note Contributor
CAS
 
Join Date: Apr 2005
Location: Hitler's Secret Bunker
Studying for ACAS
College: Hard Knocks
Favorite beer: Sam Adams Cherry Wheat
Posts: 62,056
Default

Agree with CuriousGeorge . . . is the claim count 1 or 0 for those types of claims?
__________________
I find your lack of faith disturbing

Why should I worry about dying? Itís not going to happen in my lifetime!


Freedom of speech is not a license to discourtesy

#BLACKMATTERLIVES
Reply With Quote
  #4  
Old 03-21-2018, 02:14 PM
BassFreq's Avatar
BassFreq BassFreq is offline
Member
CAS
 
Join Date: Jun 2003
Location: Chicago
Studying for all eternity
Favorite beer: Duff
Posts: 1,707
Blog Entries: 2
Default

Where are you trying to go with the freq/sev?

Loss Ratio = Losses / premium = (claim count per premium $) * (losses / claim count)

You want the claim counts to cancel each other out. So, which way would you like to go?

Perhaps, try it both ways on a subset of data and measure one the hold out data which approach performs better.
__________________
If at first you don't succeed, you have one data point.
Res ipsa loquitur, sed quid in infernos dicet?
Reply With Quote
  #5  
Old 03-21-2018, 05:05 PM
RockThatScoober RockThatScoober is offline
CAS
 
Join Date: Mar 2018
Posts: 4
Default

Quote:
Originally Posted by CuriousGeorge View Post
Depends on what you did in the frequency part.
So if I include policies with zero paid claim count (and thus zero paid amount) in frequency analysis, I have to include these in the severity analysis. But then I will have severities of 0/0?
Reply With Quote
  #6  
Old 03-21-2018, 05:11 PM
Marcie's Avatar
Marcie Marcie is offline
Member
CAS
 
Join Date: Feb 2015
Posts: 7,700
Default

Zero values are difficult for some exponential family distributions to handle.
Reply With Quote
  #7  
Old 03-21-2018, 05:14 PM
BassFreq's Avatar
BassFreq BassFreq is offline
Member
CAS
 
Join Date: Jun 2003
Location: Chicago
Studying for all eternity
Favorite beer: Duff
Posts: 1,707
Blog Entries: 2
Default

Quote:
Originally Posted by RockThatScoober View Post
So if I include policies with zero paid claim count (and thus zero paid amount) in frequency analysis, I have to include these in the severity analysis. But then I will have severities of 0/0?
You have to distinguish between zero claims (that is, records that did not have any claims and therefore had no severity either), which would be excluded from a severity model, and records with claims closed without payment, which would be included in the severity model. If you're going to go that route in a GLM, you would likely want to use a Tweedie distribution.
__________________
If at first you don't succeed, you have one data point.
Res ipsa loquitur, sed quid in infernos dicet?
Reply With Quote
  #8  
Old 03-30-2018, 09:19 AM
Strider's Avatar
Strider Strider is offline
Member
CAS
 
Join Date: Dec 2012
Posts: 1,234
Default

Quote:
Originally Posted by BassFreq View Post
You have to distinguish between zero claims (that is, records that did not have any claims and therefore had no severity either), which would be excluded from a severity model
So they would be excluded from the severity model, but they would be included in the frequency model, correct?
Reply With Quote
  #9  
Old 03-30-2018, 09:39 AM
Vorian Atreides's Avatar
Vorian Atreides Vorian Atreides is online now
Wiki/Note Contributor
CAS
 
Join Date: Apr 2005
Location: Hitler's Secret Bunker
Studying for ACAS
College: Hard Knocks
Favorite beer: Sam Adams Cherry Wheat
Posts: 62,056
Default

Quote:
Originally Posted by Strider View Post
So they would be excluded from the severity model, but they would be included in the frequency model, correct?
I would hope not.
__________________
I find your lack of faith disturbing

Why should I worry about dying? Itís not going to happen in my lifetime!


Freedom of speech is not a license to discourtesy

#BLACKMATTERLIVES
Reply With Quote
  #10  
Old 03-30-2018, 12:25 PM
Strider's Avatar
Strider Strider is offline
Member
CAS
 
Join Date: Dec 2012
Posts: 1,234
Default

Quote:
Originally Posted by Vorian Atreides View Post
I would hope not.
So you don't include records with no claims?

Otherwise I don't see how you model severity for entries with no losses,
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 04:32 PM.


Powered by vBulletin®
Copyright ©2000 - 2018, 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 0.38500 seconds with 9 queries