© 2011 towers watson. all rights reserved. intelligent use of competitive analysis 2012 ratemaking...

© 2011 Towers Watson. All rights reserved.

Intelligent Use of Competitive Analysis2012 Ratemaking and Product Management Seminar

Kelleen Arquette

March 20-21, 2012

© 2011 Towers Watson. All rights reserved. Proprietary and Confidential. For Towers Watson and Towers Watson client use only.

towerswatson.com 2

Session agenda and objectives

Today, we will discuss Different industry approaches to competitive analysis Key challenges in performing quantitative competitive analysis Analysis of “on-the-street” prices

AGENDA

Competitive Market Analysis

towerswatson.com 3



towerswatson.com 4

Insurers use various approaches to competitive market analysis — we will consider advantages and disadvantages of six specific techniques

High Degree of Sophistication

Low Degree of Sophistication

CMA: Qualitative Analysis

Market Basket

Company Statistics

Competitor Rate

Changes

Agent Feedback

CMA: Quantitative

Analysis

These options are not mutually exclusive — different approaches can be used in combination

COMPETITIVE MARKET ANALYSIS

External Quotes

Addressing Challenges

towerswatson.com 5



towerswatson.com 6

Although generally more effective, advanced competitive market analysis techniques pose certain challenges

Key Challenges Company selection Credit/tier alignment Missing variables Choice of a comparative rater Particularly for homeowners, alignment of product type Validation of results

The next several pages briefly address each of these challenges

CHALLENGES


towerswatson.com 7

Selecting which competitors to include is important…and trickier than one might think

The ideal is a mix of direct competitors and industry leaders The target market segment should be considered

Competitors targeting the preferred market may be different than competitors targeting the non-standard market

Once you choose a competitor group, selecting which particular company to rate can be challenging For example, Allstate writes auto insurance in at least 14 companies across the country

Relative premium volume may not be a good indicator, as one entity may be a new company (where all new business is being written)

Agent feedback and rate filing reviews can help select the “right” company Some companies write only package policies (personal auto and homeowners on the

same policy). This should be considered in the company selection (impact on coverage alignment and underwriting selection criteria)

COMPANY SELECTION


towerswatson.com 8

Creating an accurate alignment between competitor credit groups and tiers is critically important

If a comparative rater provides automatic credit scoring alignment at all, it usually provides a simplified approach, such as this:

Credit Group

A

B

C

D

Credit Group

1

2

Company A Company B Alignment

Company A

Company B

A, B 1

C, D 2

This approach has obvious disadvantages and could lead to incorrect conclusions

CREDIT/TIER ALIGNMENT


towerswatson.com 9

Alternative approaches to aligning credit groups/tiers can increase accuracy (but can be costly and/or time consuming)

Insurance Score Alignment (Distribution Mapping) – Alignment based on company filed distributions by credit score range or tier Relies on publicly available data More accurate than uniform distribution assumption

Insurance Score Assignment – Assignment based on programmed competitor credit scoring algorithms Requires data at the individual credit attribute level Relies on publicly available data Processing current book of business through programmed algorithms results in an optimal credit

score assignment for each competitor Assumptions may still be necessary, depending on the data source and competitor(s)



towerswatson.com 10

Competitor 3 InsScore x Competitor 4 InsScore

60

70

80

90

100

110

120

130

140

150

160

170

180

190

450 500 550 600 650 700 750 800

Competitor 3 InsScore

Co

mp

eti

tor

4 In

sS

co

re

Credit-based insurance score used by different companies assess risk differently

“Company A” and “Company B” are personal auto insurers Both are national writers with

market share in the top 10 in most states

Credit-based insurance scoring models Company A uses a vendor model

— High score is best (lowest risk)

Company B uses a proprietary model— Low score is best (lowest risk)

Models were found in publicly available filings

Models were programmed using actual credit data

Correlation between the insurance scores, but not perfect

Expect diagonal line if models assessed risk in the same way

No hits/no scores are excluded Company A Insurance ScoreC

om

pan

y B In

suran

ce Sco

re

Company A X Company B Insurance Score Contour Plot



towerswatson.com 11

Companies take different approaches to tier

Tier is a combination of the credit-based insurance score and other variables for both companies

Company A and Company B use different variables in the tier determination

The same data set was used to generate both tier graphs

Examples of variables used include Prior liability limits Lapses in coverage Education Occupation Accident and violations Length of time insured with prior carrier


Company A Tier

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Tier

Ve

hic

le C

ou

nt

Company B Tier

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Tier

Ve

hic

le C

ou

nt


towerswatson.com 12

It is possible for a policy considered low risk for Company A to be considered high risk for Company B

Any tier for Company A has a range of tiers for Company B

Can explain pricing differences at the individual vehicle/policy level

Insurance score or tier alignment approaches miss the opportunity to look at the different approaches to risk assessment at the policy level

Company A Tier x Company B Tier

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Company A Tier

Dis

trib

uti

on

Company B Tier 18

Company B Tier 17

Company B Tier 16

Company B Tier 15

Company B Tier 14

Company B Tier 13

Company B Tier 12

Company B Tier 11

Company B Tier 10

Company B Tier 9

Company B Tier 8

Company B Tier 7

Company B Tier 6

Company B Tier 5

Company B Tier 4

Company B Tier 3

Company B Tier 2

Company B Tier 1


Company A Tier x Company B Tier

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Company A Tier

Ve

hic

le C

ou

nt

Company B Tier 18

Company B Tier 17

Company B Tier 16

Company B Tier 15

Company B Tier 14

Company B Tier 13

Company B Tier 12

Company B Tier 11

Company B Tier 10

Company B Tier 9

Company B Tier 8

Company B Tier 7

Company B Tier 6

Company B Tier 5

Company B Tier 4

Company B Tier 3

Company B Tier 2

Company B Tier 1


towerswatson.com 13

In some cases, a company may simply not collect accurate data on a rating variable that a competitor uses

Depending on the importance of the variable, how missing values are populated can materially affect the results

External data can sometimes be used to fill in missing values using sampling techniques Census and other external data Distributions obtained from competitor filings

Care should be taken in how these variables are populated Suppose a company does not collect data on a 55 & Retired Discount, but driver age is readily

available From census data and other publicly available data, one can obtain a population estimate of

individuals who are retired However, constraints should be placed on the sampling approach to avoid illogical results (e.g.,

a 25-year-old who is “retired”)

MISSING VARIABLES


towerswatson.com 14

There are a number of important due diligence considerations in selecting a comparative rater

Are the rates for your company and the selected competitors already included in the tool (and if not, what are the additional costs to include?) What are the alternatives if additional cost is prohibitive?

Does the software support batch rating hundreds of thousands of policies (or more) in a timely fashion? How much computing power is necessary?

Does the platform accurately perform: Driver assignment for personal auto? Territorial assignment? Tier assignment?

What process does the vendor have in place to ensure accurate results? What types of training and support services does the vendor provide? Does the vendor have a tool to convert your exposure data into a format that the batch

rater can use? Does the vendor have appropriate marketplace knowledge to understand complex rate

filings?

Although companies can decide to perform this work in-house, the effort has significant staffing and cost implications

COMPARATIVE RATER


towerswatson.com 15

Proper alignment of product/coverage is important in order to draw appropriate conclusions

Coverage/FeatureCompetitor A

“Standard” HO-3 PolicyCompetitor B

“Basic” HO-3 Policy

Earthquake Included Excluded

Water Backup Excluded Included

Coverage A Actual cash value, with possible limited replacement cost coverage endorsement

Replacement cost coverage

Coverage C 70% of Coverage A 85% of Coverage A

Identity Theft Included Excluded

PRODUCT ALIGNMENT

State X — Homeowners


towerswatson.com 16

What process will be followed to ensure that the calculated competitor rates are accurate?

Rater accuracy should be considered in selecting a third-party vendor Even the larger comparative rating vendors are often not accurate

Programming errors Credit/tiering alignment Oversimplification/misunderstanding of a competitor’s rating approach Goal may be to get “close enough”

Certain actions can be taken to increase the accuracy of the analysis Hand-rating of a random sampling of policies (which can be time consuming) Verifying calculated average premiums with certain filed materials Conversations with agents (“gut checks”)

VALIDATION OF RESULTS


towerswatson.com 17

Analysis of “On-The-Street” Prices

towerswatson.com 17



towerswatson.com 18

ILLUSTRATIVE

Average Premium All Coverages Combined

Company A Company B Company C

Company D

Company E Company F

Company G

XYZ Company

$0

$200

$400

$600

$800

$1,000

$1,200

$1,400

Average Premium

Beware of potential inherent bias in using current policy mix of business

QUANTITATIVE ANALYSIS

A quantitative analysis can compare new and renewal business pricing against competitors for the entire book…


towerswatson.com 19

$500

$700

$900

$1,100

$1,300

$1,500

$1,700

1 2 3 4 5 6 7 8 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

0

500

1,000

1,500

2,000

2,500

3,000

3,500

Policy Count Company Premium Second Highest Premium Second Lowest Premium


Average Premium with Hi/Lo Band by Symbol Policy

Count

Symbol

ILLUSTRATIVE

Average Premium

…and by rating factor/segment…


towerswatson.com 20

The target market position should be identified and then metrics can be developed to monitor competitive position relative to target

…to identify the current competitive position


Competitive Metrics $ or % Competiveness % Wins Relative to Market Rank Average premiums

Other metrics may also be considered when determining the target market position $ or % Impacts $ or % Subsidization Expected loss ratio Expected retention Hit/conversion ratio


towerswatson.com 21

Now what?

Identified current competitive position Calculated current premiums for your company and selected competitors, overcoming many hurdles

in calculating the competitor premiums Analyzed the premiums to determine the current competitive position

Identified target competitive position How do you get there?

How should the current rating plan be revised to achieve the target competitive position?



towerswatson.com 22

Contact Information

Kelleen Arquette Consultant, Towers Watson [email protected] 509-258-8876

CONTACT INFORMATION

© 2011 towers watson. all rights reserved. intelligent use of competitive analysis 2012 ratemaking...

Documents