Transcript Deloitte Consulting, 2005 © Deloitte Consulting, 2005

© Deloitte Consulting, 2005
What To Do When You Cannot Use
Credit? (Personal Lines)
Cheng-Sheng Peter Wu, FCAS, ASA, MAAA
CAS 2005 Special Interest Seminar
Chicago
September 19-20, 2005
© Deloitte Consulting, 2005
Agenda
 The credit scoring revolution
 What to do when cannot use credit?
 Conclusions
2
© Deloitte Consulting, 2005
The Credit Score Revolution
© Deloitte Consulting, 2005
Personal Lines Pricing and Class
Plans – History



Few rating factors before World War II
Explosion of class plan factors after the War
Auto class plans:


Homeowners class plans:


Territory, driver, vehicle, coverage, loss and violation, others,
tiers/company…
Territory, construction class, protection class, coverage, prior
loss, others, tiers/company...
Credit scoring introduced in late 80s and early 90s
4
© Deloitte Consulting, 2005
Personal Lines Credit Scoring –
History





First important factor identified over the past 2
decades
Composite multivariate score vs. raw credit
information
Introduced in late 80s and early 90s
Viewed at first as a “secret weapon”
Quiet, confidential, controversial, black box, …etc
“Early believers and users have gained
significant competitive advantage!”
5
© Deloitte Consulting, 2005
Early Believers’ Benefit from Credit
Scores
0.35
115%
0.3
110%
0.25
105%
0.2
100%
0.15
95%
0.1
90%
0.05
85%
0
Combined Ratio
Growth Rate
Progressive vs Industry
Industry Growth Rate
Progressive Growth Rate
Industry Combined Ratio
Progressive Combined Ratio
80%
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
Year
6
© Deloitte Consulting, 2005
The Current Environment

Now everyone is using it:




Marketing and direct solicitation
New business and renewal business pricing and underwriting
How to stay competitive if everyone is using it?
Regulatory constraints:



Many states have conducted studies on the true correlation
with loss ratio and potential discrimination issues - WA study,
TX study, MO study
Many states have/are considering restricting the use of credit
scores or certain types of credit information
More states want the “black box” filed and opened
7
© Deloitte Consulting, 2005
Some Facts About Credit Scores



A composite score that usually contains 10 to 40
pieces of credit information
Loss ratio lift is significant – a powerful class plan
factor or rate tiering factor (2.0 ratio of worst 10 to
best 10%)
Benefits/ROI are measurable


Lift curve can be translated into bottom-line benefit
Blind test and independent validation can be done to
verify the benefit
8
© Deloitte Consulting, 2005
Loss Ratio Lift Curve
120
90
82
78
Loss Ratio
74
66
70
62
58
50
Credit Score Decile
9
© Deloitte Consulting, 2005
Credit Score Revolution Segmentation Power
1997 NCCI/Tillinghast Study of 9 Companies' Data
Loss Ratio Relativity of the Best and Worst 20% of Credit Score
Co1
Co2
Co3
Co4
Co5
Co6
Co7
Co8
Co9
Avg
Best 20%
-38% -29% -19% -15% -14% -34% -22% -22% -36% -25%
Worst 20%
48%
20%
32%
30%
46%
59%
20%
22%
95%
41%
10
© Deloitte Consulting, 2005
What to Do when Cannot Use
Credit Scores?
© Deloitte Consulting, 2005
What to Do when Cannot Use
Credit

One idea is to find “Credit Score Proxies”
“Length of account” --- “Length of policies”, “Age
of policyholders”?
 “Late payment” --- “Late payment in paying
premium bill”, “Insurance lapse”?
 “Derogatory / Bankruptcy information etc” --- who
has less chance to have derogatory or bankruptcy?
 etc…

12
© Deloitte Consulting, 2005
What to Do when Cannot Use
Credit

Another idea is - why limited to “Credit
Proxies” only, and go from credit scores to
data mining and predictive modeling
A credit score is just one example of an insurance
predictive model
 The same methods used to build credit scores are
used in data mining to build insurance predictive
models – “Go Beyond Credit Models”.
 Broaden the usage of “predictive variables”

13
© Deloitte Consulting, 2005
Go Beyond Credit Models

The key is to use as much information as
possible
 in

a multivariate way
Choice of statistical techniques is important,
but the real key is the quality and breadth of
predictive variables used.
 GIGO
 Actuarial/insurance

knowledge is critical
Untapped riches reside in many companies’
transactional records.
14
© Deloitte Consulting, 2005
Data Sources


We classify possible data sources into two
groups
Internal data sources: predictive information
gleaned from the company’s own systems
 Regardless

of how or whether it is currently used
External data sources: predictive information
available from 3rd parties.
 Both
credit and non-credit
15
© Deloitte Consulting, 2005
Internal Data Sources

Policy information
 Limits,
Deductibles, Measure of exposure (# cars,
#houses, #employees, $sales, premium size…

Line-Specific information
 Driver,

Vehicle, Business Class …
Policyholder information
 Age,
gender, marital status …
16
© Deloitte Consulting, 2005
Internal Data Sources


Customer-level information
Transactional data
 Coverage,

premium and loss transactions
Billing information
 Correlation
with credit
Agent information
A little creativity in using these data sources will
go a long way!

17
© Deloitte Consulting, 2005
An Example of a “Creative” Variable

“Distance between Agent and Insured”: close
by agents know you better!
 Insured’s
address available in policy data system
 Agent’s address available in agency database
 Map two addresses into longitude and latitude using
“geo-coding” tools
 Calculate the distance using “longitude-latitude distance
formula”
18
© Deloitte Consulting, 2005
External Data Sources

Credit
 Predictive


both for commercial and personal lines
MVR – CLUE
Zipcode/geographic information
 Rating
territory
 Many different sources available

The sky is the limit but
 Consider
cost, hit rate, implementation, …etc
19
© Deloitte Consulting, 2005
Types of Variables Generated

Territory-level
 Demographic,

Policy / policyholder-specific
 Many

weather, crime, ...etc
traditional rating variables fall into this category
Behavioral
traditional – fits more neatly into data mining
paradigm than classification ratemaking
 Credit, billing, prior claims, cancel-reinstatements…
 Less
20
© Deloitte Consulting, 2005
How Many Variables?

It is possible to generate literally hundreds of
predictive variables
Some will be redundant
 Some will not be very predictive
 Some will be somewhat predictive
 Some will be “killer”


A good model can contain as few as 15-20 or
as many as 60-70 variables

Usually no single “ideal” model
21
© Deloitte Consulting, 2005
Which Variables to Use?

Choosing is a major part of the data mining process



Use variety of exploratory statistical techniques
Use prior modeling experience / actuarial knowledge
Several considerations





Actuarial / underwriting knowledge
Client’s business needs
Legal / regulatory considerations
Data availability / cost
Systems implementation considerations
22
© Deloitte Consulting, 2005
In Our Experience….
Do “Go-Beyond Credit” PMs work?

YES: non-credit predictive models are
Valuable alternative to credit scores
 Flexible
 Tailored to individual companies
 Leverage company’s untapped internal data
 Comparable predictive power to credit scores


And mixed credit / non-credit PMs can be
even stronger
23
© Deloitte Consulting, 2005
…But It’s Not a Walk Through the
Park
Challenges for PMs:





IT resources constraints
Project management
Business process buy-in
Success of system and business implementation
Training and organizational change
24
© Deloitte Consulting, 2005
Conclusions
© Deloitte Consulting, 2005
Industry Trends



How do companies try to stay competitive regarding the
use of credit?
How do companies prepare for increasing regulatory
constraints?
Industry trends



Companies are developing modeling capabilities and
pursuing various applications
Companies are developing proprietary credit scoring models
rather than buying “off-the-shelf” credit scores.
Companies are also going beyond credit, to build scoring
models that don’t rely solely on credit
26
© Deloitte Consulting, 2005
Keys to Building Credit Alternative
Models

Fully utilize all sources of information
 Leverage
company’s internal data sources
 Enriched with other external data sources





Use large amount of data
Employ systematic analytical process
Use state-of-the-art modeling tools
Apply multivariate methodology
Disciplined project management
27