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 03-17-2017, 04:14 PM
whoanonstop's Avatar
whoanonstop whoanonstop is offline
Member
Non-Actuary
 
Join Date: Aug 2013
Location: Los Angeles, CA
Studying for Spark / Scala
College: College of William and Mary
Favorite beer: Orange Juice
Posts: 5,717
Blog Entries: 1
Default Predictive Prudence

Read through this actuarial review article.

Just my observations, good, bad and ugly.

Good (I agree with)

Quote:
However, since insurance companies tend to have a mindset to improve financial results in the short term, she said, it is difficult to get them thinking about innovation “that will not pay off in six months.”
Definitely agree with this. I find it ironic since it feels like insurance companies should be posed for the long term, but individual stakeholders within the company look for short-term payoffs because they usually affect their bonuses directly.

Quote:
Regulatory restrictions — whether real or perceived — can also hamper innovation
Definitely.

Quote:
The advantage of applying analytics for decision-making is that the techniques provide an objective anchor...

Without it, he said, “the best an organization can do is to have an average performance that is a function of the independent aggregated thinking of every person.”
This is the best statement in the entire article. It would be interesting to see the total cost of certain pet theories people have carried for years to any industry that doesn't objectively anchor.

Quote:
“Data is becoming more important than business relationships or clinical knowledge,” Lowe observed.
Related to previous statement. Perfect.

Bad (I disagree with)

Quote:
Many obstacles stem from the cautious nature of insurance companies.
I do agree that insurance companies are cautious, but I think the main reason why they are slow to adopt any new methods is either due to regulatory framework or a lack of knowledge in those new methods.

Quote:
Insurance companies often prefer waiting to see if an approach is tested and proven, and if it will impact the bottom line, Mosley said.
Companies would also prefer to wait if they don't have the manpower or ability to do something, no? Similar statement as previous.

Quote:
the competitive edge to be gained from predictive modeling innovation can be short-lived in this “quick-to-copy” industry.
I hate this statement. Mostly because I think it is a weak excuse to not do better. Do you think Uber was like... You know... we have this idea... but what happens if someone copies us later? Plus, the higher the difficulty level, the harder to copy.

Quote:
Actuaries can learn data science techniques and data scientists can gain deeper industry knowledge through the iCAS program.
I'm not feeling this, but maybe I'll be proven wrong.

Ugly (wtf)

Quote:
“The same companies that resisted GLMs 15 years ago,” Mosley said, “are now saying the same things to me about advanced modeling."
What does "advanced modeling" entail? Not a single reference in the article as to what this means.

Quote:
One challenge of implementing advanced machine-learning models is that they can appear as black boxes to regulators
Again, what does "advanced machine-learning" mean? Plus, I think it is generally a good idea that if you're someone who builds models to never refer to them as "black boxes". I think that gives a bad perception of newer methods.

Quote:
Steve Lowe, a senior consultant with Willis Towers Watson, said that the transition from the traditional model to one that is data-driven often begins with combining actuaries and data scientists on innovation teams.
Does relabelling actuaries as Data Scientists count?

Quote:
Big Data
Quote:
big data
Quote:
big data
Quote:
big data
Quote:
big data
Quote:
big data
Quote:
Big Data
Quote:
big data
Quote:
Big Data
Stahp.

-Riley
__________________
It is impossible to have a professional forum where the majority of your professionals are anonymous.

Map of Actuarial Hiring Companies
Reply With Quote
  #2  
Old 03-17-2017, 04:24 PM
Colonel Smoothie's Avatar
Colonel Smoothie Colonel Smoothie is online now
Member
CAS
 
Join Date: Sep 2010
College: Jamba Juice University
Favorite beer: AO Amber Ale
Posts: 46,539
Default

I thought this was going to be about Dear Prudence
__________________
Recommended Readings for the EL Actuary || Recommended Readings for the EB Actuary

Quote:
Originally Posted by Wigmeister General View Post
Don't you even think about sending me your resume. I'll turn it into an origami boulder and return it to you.
Reply With Quote
  #3  
Old 03-17-2017, 04:38 PM
Colonel Smoothie's Avatar
Colonel Smoothie Colonel Smoothie is online now
Member
CAS
 
Join Date: Sep 2010
College: Jamba Juice University
Favorite beer: AO Amber Ale
Posts: 46,539
Default

Quote:
Originally Posted by whoanonstop View Post
What does "advanced modeling" entail? Not a single reference in the article as to what this means.
I'm sure you could email him for further clarification. His info is in the directory.
__________________
Recommended Readings for the EL Actuary || Recommended Readings for the EB Actuary

Quote:
Originally Posted by Wigmeister General View Post
Don't you even think about sending me your resume. I'll turn it into an origami boulder and return it to you.

Last edited by Colonel Smoothie; 03-17-2017 at 04:52 PM..
Reply With Quote
  #4  
Old 03-17-2017, 04:48 PM
Colonel Smoothie's Avatar
Colonel Smoothie Colonel Smoothie is online now
Member
CAS
 
Join Date: Sep 2010
College: Jamba Juice University
Favorite beer: AO Amber Ale
Posts: 46,539
Default

Quote:
Originally Posted by whoanonstop View Post
Again, what does "advanced machine-learning" mean? Plus, I think it is generally a good idea that if you're someone who builds models to never refer to them as "black boxes". I think that gives a bad perception of newer methods.
Did you face this problem when you worked in your non-actuarial jobs? I would be surprised if you haven't. I agree that you shouldn't refer to them as black-boxes. Especially something as simple as GLM. A long time ago, an actuary I worked with deliberately tried to avoid some very-appropriate questions from an underwriter by characterizing a rating algorithm as a black-box that they wouldn't understand. I think you can guess how that turned out.

But even when you're not referring to something as a black-box, doesn't mean your stakeholders won't. What if your best staff leave because you're uncompromising on letting them do their jobs?
__________________
Recommended Readings for the EL Actuary || Recommended Readings for the EB Actuary

Quote:
Originally Posted by Wigmeister General View Post
Don't you even think about sending me your resume. I'll turn it into an origami boulder and return it to you.
Reply With Quote
  #5  
Old 03-17-2017, 04:51 PM
MoralHazard MoralHazard is offline
Member
CAS
 
Join Date: Jul 2011
Favorite beer: Sam Adams Rebel Rouser
Posts: 93
Default

Quote:
Originally Posted by whoanonstop View Post
What does "advanced modeling" entail? Not a single reference in the article as to what this means.
-Riley
I thought it was pretty clear -- methods other than GLMs or GLM-like models, such as GBM/Random Forest/neural net/whatever where you get enhanced predictive power but no clean interpretation like you do with GLMs. The article was written for a broad audience, hence the lack of technical specifics.
Reply With Quote
  #6  
Old 03-17-2017, 04:58 PM
r. mutt's Avatar
r. mutt r. mutt is offline
Member
 
Join Date: May 2002
Location: Extreme fever swamps of the left
Posts: 10,130
Default

Quote:
Originally Posted by Colonel Smoothie View Post
I thought this was going to be about Dear Prudence
Is that why you came out to play?
__________________
(\ (\
( ^_^)
(_(")(") It all comes full akchigs eventually. ~ Full-On Devi
Reply With Quote
  #7  
Old 03-17-2017, 05:30 PM
whoanonstop's Avatar
whoanonstop whoanonstop is offline
Member
Non-Actuary
 
Join Date: Aug 2013
Location: Los Angeles, CA
Studying for Spark / Scala
College: College of William and Mary
Favorite beer: Orange Juice
Posts: 5,717
Blog Entries: 1
Default

Quote:
Originally Posted by MoralHazard View Post
I thought it was pretty clear -- methods other than GLMs or GLM-like models, such as GBM/Random Forest/neural net/whatever where you get enhanced predictive power but no clean interpretation like you do with GLMs. The article was written for a broad audience, hence the lack of technical specifics.
I did not know that the Actuarial Review is written for a broad audience. My problem with the word "advanced modeling" is that it is extremely subjective. If they would have said "advanced modeling, such as XXXXX" that wouldn't hurt the audience and would better clarify what they are talking about. Right now "advanced modeling" is a catch-all statement that doesn't boost confidence that the writer knows what they are talking about (that could or could not be the case). As long as there are no full out explanations of what something like Deep Learning entails, then the audience focus still remains broad.

(Advanced modeling is equivalent to the term "Big Data" in my mind)

I don't agree that GLMs always give a good interpretation, but they are perceived as always reliable and make people who don't know anything about modeling feel overly comfortable.

-Riley
__________________
It is impossible to have a professional forum where the majority of your professionals are anonymous.

Map of Actuarial Hiring Companies
Reply With Quote
  #8  
Old 03-17-2017, 05:43 PM
Marcie's Avatar
Marcie Marcie is online now
Member
CAS
 
Join Date: Feb 2015
Posts: 6,824
Default

Riley, I'm not sure I understand your objections to the statements in your "bad" category, especially the first two. You disagree with reasons because you think there are additional reasons? Do you reject all models that don't include every possible explanatory variable?

I agree that the third one is a weak excuse for not doing better, but it's also a business reality that your R&D edge can disappear quickly with a "me too" filing from one or more competitors, meaning you might not be able to get an appropriate return on R&D investment. That's tough to quantify, and like I said, I agree with you on this one. It also suggests that advanced techniques (whatever that means) might be happening but kept for in-house applications (e.g. underwriting, claims, marketing) rather than publicly-filed rate plans.

The fourth is probably much weaker than the marketing folks at iCAS would like to admit, but it would be hard for it to be false.
Reply With Quote
  #9  
Old 03-17-2017, 05:53 PM
Colonel Smoothie's Avatar
Colonel Smoothie Colonel Smoothie is online now
Member
CAS
 
Join Date: Sep 2010
College: Jamba Juice University
Favorite beer: AO Amber Ale
Posts: 46,539
Default

first-mover advantage?
__________________
Recommended Readings for the EL Actuary || Recommended Readings for the EB Actuary

Quote:
Originally Posted by Wigmeister General View Post
Don't you even think about sending me your resume. I'll turn it into an origami boulder and return it to you.
Reply With Quote
  #10  
Old 03-17-2017, 05:59 PM
whoanonstop's Avatar
whoanonstop whoanonstop is offline
Member
Non-Actuary
 
Join Date: Aug 2013
Location: Los Angeles, CA
Studying for Spark / Scala
College: College of William and Mary
Favorite beer: Orange Juice
Posts: 5,717
Blog Entries: 1
Default

Quote:
Originally Posted by Marcie View Post
You disagree with reasons because you think there are additional reasons? Do you reject all models that don't include every possible explanatory variable?
All I'm saying is the fingers are being pointed at one reason because the field struggles to be humble and say... "Look, we don't have our shit together to do this stuff". Some companies might.

Not sure if the second question is meant to connect to some other statement that I can't see, but I actually don't believe every model should be "complex". I'm saying there are a select few that would definitely be beneficial if they were.

Quote:
Originally Posted by Marcie View Post
It also suggests that advanced techniques (whatever that means) might be happening but kept for in-house applications (e.g. underwriting, claims, marketing) rather than publicly-filed rate plans.
Unless something changes with the rate filing procedure, I definitely agree that the best opportunities are in claims/underwriting/marketing. Lots of potential to be more efficient there.

Quote:
Originally Posted by Marcie View Post
The fourth is probably much weaker than the marketing folks at iCAS would like to admit, but it would be hard for it to be false.
Personally, I've yet to see evidence that would suggest this credential is going to do well and I would absolutely love to see it do well. I'm biased towards it because I want this stuff to be more common knowledge.

-Riley
__________________
It is impossible to have a professional forum where the majority of your professionals are anonymous.

Map of Actuarial Hiring Companies
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 02:25 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 0.31213 seconds with 7 queries