Data Analytics Tuesday, September 29, 2015

Size: px

Start display at page:

Download "Data Analytics Tuesday, September 29, 2015"

Osborne Mills
5 years ago
Views:

1 Data Analytics Tuesday, September 29, 2015 Jeff Kucera, FCAS, MAAA Actuary and Consultant e2value, Inc. Hawthorn Woods, Ill. Jeff Kucera is an actuary and consultant for e2value, the leading provider of web-based insurance-to-value software. He is a member of the board of trustees of MSO, Inc., a rating and statistical bureau located in Glen Rock, N.J. Previously, Jeff spent approximately 25 years with Allstate Insurance and 12 years as an actuarial consultant. Jeff has spent the majority of his career working in personal lines, particularly the pricing and marketing of the homeowners and private passenger auto segments. Brett Nunes Director, Risk Consulting and Software Practice Towers Watson Chicago, Ill. Brett Nunes is a director in the risk consulting and software practice of Towers Watson, a leading global professional services company that helps organizations improve performance through effective management of people, risk, and finances. Brett has been a consultant in the insurance industry for more than 15 years. He has primarily served personal lines and commercial lines carrier clients by using analytics to study and solve various business needs through improvement of rating, underwriting, marketing, and claims processes.

2 Three Session Ideas Tools or tips you learned from this session and can apply back at the office

3 Insurance Analytics for Small and Medium Insurers Agenda Data that Companies Are Reviewing Case Study New Homeowner Rating Plan Structure Insurance Score Conclusion 2015 Connect Differently - Kucera & Nunes Page 1 of 11

4 Data that Companies Are Reviewing Private Passenger Auto Home Ownership Payment Method Time Car Owned Years at Address Garage Zip Different from Mailing Zip Gender of Oldest Operator Data that Companies Are Reviewing Private Passenger Auto Usage Based Insurance Number of Hard Brakes Frequency of Lane Switches Time of Day Car Driven Hard Starts Difference Between Driving Speed and Speed Limit 2015 Connect Differently - Kucera & Nunes Page 2 of 11

5 Data that Companies Are Reviewing Homeowners Insurance Auto Losses Auto Violations Horsepower of Car Method that Insurance Bill Is Paid Days Between Application and Effective Date Data that Companies Are Reviewing Homeowners Insurance Original Owner of Home Prior Renter/Condominium Insured Age/Type of Roof Number of Bathrooms Number of Stories 2015 Connect Differently - Kucera & Nunes Page 3 of 11

6 Case Study New Homeowner Rating Plan Background Small company writing homeowners in only two states Not large enough to do GLM Had not introduced rating for credit or prior losses (nor many other recent rating plans) Growth was negative and profitability was near 0% Case Study New Homeowner Rating Plan Company Objectives Primary Profitable Growth Secondary Revised Homeowner Rating Plan Introduce More Sophisticated Rating Plan WITHOUT GLMs 2015 Connect Differently - Kucera & Nunes Page 4 of 11

7 Case Study New Homeowner Rating Plan Steps Taken by Company and Consultant Acquired Manual of Top Ten Writer for Review Reviewed Manual and Competitor Information to Develop Wish List of Rating Characteristics Developed Methods to Reasonably Estimate Rating Factors Case Study New Homeowner Rating Plan Issues that Needed to Be Addressed How to convert by-peril rating to options that could be programmed by the company Decided to use more than one rating group for key rating characteristics Created five rating groups for key rating characteristics As examples, formed five amount of insurance scales, five deductible relativities, etc Connect Differently - Kucera & Nunes Page 5 of 11

8 Case Study New Homeowner Rating Plan Issues that Needed to Be Addressed How to assign each territory/zip code to the appropriate rating group Used analytics to review data in manual and estimate peril distribution for each zip code/territory Reviewed peril distributions in order to select peril distribution for each of the rating groups Assigned each territory/zip code to the rating group with the most similar peril distribution Case Study New Homeowner Rating Plan Steps Taken by Company Communicated how they expected new business to be written Created dashboards to monitor results monthly and look for discrepancies Made modifications when appropriate 2015 Connect Differently - Kucera & Nunes Page 6 of 11

9 Case Study New Homeowner Rating Plan Results for Company Averaged nearly 10% growth since implementation Combined ratio has consistently been well below 90% Structure Insurance Score: Background Many companies are already using some house characteristics in rating of homeowners and dwelling fire Some examples include number of bathrooms, number of stories and age of roof A greater use of more home characteristics and/or a structure insurance score will improve rating even more 2015 Connect Differently - Kucera & Nunes Page 7 of 11

10 Structure Insurance Score: Hypotheses Characteristics used to determine replacement cost above and beyond traditional dwelling or building information that carriers use to rate and underwrite property coverage may add significant lift or improvement in segmentation Level of lift will vary by peril Dwellings with identical characteristics will vary in lift due to geographic location and peril mix Similar in concept to make/model rating for automobile Structure Insurance Score: What Is It? Measures the inherent risk to an insurance company based on several different characteristics of a home or building Based on predictive modeling applications Recognizes specific characteristics of the home or building Applicable to residential, farm and commercial type risks 2015 Connect Differently - Kucera & Nunes Page 8 of 11

Structure Insurance Score: Residential Example 320 Secor Dr.

14 year old home $340,000 Replacement Cost Protection Class 6 Structure Insurance Score: Residential

Stories Square Footage Construction 145 All Other Square Footage Zip Code # Baths 140 110 Liability

11 Structure Insurance Score: Residential Example 320 Secor Dr. 14 year old home $340,000 Replacement Cost Protection Class Secor Dr. 14 year old home $340,000 Replacement Cost Protection Class 6 Structure Insurance Score: Residential Example Theft Fire Water Foundation # Baths Age of Plumbing # Stories Age of Home # Entrances 170 # Stories Square Footage Construction 145 All Other Square Footage Zip Code # Baths Liability Outbuilding Square Footage Swimming Pool Secor Dr. Weather Age of Home Roof Material Weather Mitigation Connect Differently - Kucera & Nunes Page 9 of 11

12 Structure Insurance Score: Residential Example Theft Fire Water Sprinklers Square Footage Age of Home Protective Device Zip Code Detached Garage 70 Population Density Age of Roof Sprinklers 60 All Other Square Footage Non-seasonal Residence Age of Furnace/AC Liability Zip Code # Stories Population Density 95 Weather Square Footage Type of Home Weather Mitigation Secor Dr. Structure Insurance Score: Anticipated Outcomes Hypothetical Loss Ratio by Decile 50% 100% 45% 40% 90% 35% 80% Exposure 30% 25% 20% 15% 3x Lift 70% 60% 50% Loss Ratio 10% 5% 40% 0% 30% 2015 Connect Differently - Kucera & Nunes Page 10 of 11

13 Scores are above and beyond traditional factors 1400 Score vs Premium 1200 Premium These risks are charged the same premium, but risk 1 is 3 times riskier than risk Score Conclusions Even the smallest companies can track results Beneficial to assign dedicated resource(s) to track and help interpret results Outside expertise may be necessary to get started and for larger projects 2015 Connect Differently - Kucera & Nunes Page 11 of 11

Data Sources. Rob Curry Roosevelt Mosley National Association of Insurance Commissioners

Data Sources. Rob Curry Roosevelt Mosley National Association of Insurance Commissioners Data Sources Rob Curry Roosevelt Mosley 2017 National Association of Insurance Commissioners DATA SOURCES Rob Curry, MAAA, FCAS Roosevelt Mosley, MAAA, FCAS Learning Objectives At the end of this presentation,