Accelerated Underwriting

Transcription

1 Accelerated Underwriting Derek Kueker, FSA, MAAA Vice President and Sr. Actuary, Data Solutions, RGAx May 24, 2017

2 Customer s Ideal Insurance Journey Jenny and Steve just had their third child. She works part-time and doesn t have any life insurance. He has a group life policy at work and an individual policy. She qualifies for accelerated underwriting and signs her application just days later thinking, Why did I put this off so long? Now I can stop worrying. Jenny says, I don t have life insurance. What if something were to happen? Between the new baby, little league and work, the conversation wanes. Jenny s agent contacts her to say she may qualify for an accelerated program that doesn t require an exam, and her premium will not be higher. Jenny works with the agent to submit an application. 2

3 My Personal Insurance Journey (2017) To buy a life insurance policy Double my life insurance should be EASY Better be EASY My policy is only 2 years old should be EASY I ll call my guy and it should be that EASY 3

4 My Personal Insurance Journey (2017) To buy a life insurance policy Reality Call my guy Full Needs Analysis Complete Application Do I have 1 policy, 2 policies, none? 8 weeks still no word (2 premiums paid) Frustrated Annoyed Confused Ready to quit Concerned NOT SURPRISED Agent call your doctor to expedite reports (twice) 6 weeks no word Nurse + Needles + Scheduling Pay Premiums 4

5 The Market Today 5

6 Today s Environment 58 Million households lack adequate life insurance coverage, resulting in a coverage gap of over $16 trillion. The Life insurance industry is trying to mitigate the underwriting issues that contribute to this gap by: TRL = 45 Placing a greater importance on consumer experience Improving process speed and efficiency Embracing a data driven process Increasing transparency Improving consumer trust SOURCE: LIMRA Life Insurance, Why So Down and LIMRA Life Insurance Coverage Gap Substantial and Growing 6

7 Today s Environment Where are we going? Consumers expect immediate satisfaction But Faster processes lead to more anti-selection So The industry is balancing experience with product design And TRL = 45 Data driven solutions are bridging the gap How will the market react? 7

8 What Accelerated Underwriting Enables You To Do Identify applicants who qualify for fluidless underwriting Issue near-fully underwritten retail rates 8

9 Accelerated Underwriting Helps Balance Market Demands Affordability Better process Applicants want Carriers want Reliable underwriting evidence New forms of evidence and advanced data analysis techniques are starting to balance the needs of both groups 9

10 Accelerated Underwriting Challenge of balancing business requirements Percentage of Accelerated Applicants Mortality Slippage Age Qualifications Face Amount Qualifications Expense Savings Changes in Take-up Rates Retail Rates Agent Communications The Unknown 10

11 Accelerated Models Can Vary Different designs and approaches in the market Most target elimination of some underwriting elements in the age/amount grid Can result in a wide range of mortality outcomes 11

12 Accelerated Underwriting A Dynamic Process Application Full Application with Tele-Interview & Drill Downs Accelerated / Fluidless Path Low TRL Scores Issue Age & Face Amount Limitations Y Gather 3 rd Party Data Meets Req s Y Underwrite & Make Offer without additional testing Audit N N Initial Screen TrueRisk Life Score Intermediate TRL Scores Apply Full Underwriting High TRL Scores Order Additional Req s 12

13 Accelerated Underwriting A Dynamic Process Application Full Application with Tele-Interview & Drill Downs Low TRL Scores Initial Screen TrueRisk Life Score High TRL Scores Issue Age & Face Amount Limitations Y Gather 3 rd Party Data Meets Req s Y Underwrite & Make Offer without additional testing Audit N N Intermediate TRL Scores Apply Full Underwriting Traditional Full Underwriting Order Additional Req s 13

14 Accelerated Underwriting A Dynamic Process Application Full Application with Tele-Interview & Drill Downs Low TRL Scores Issue Age & Face Amount Limitations Y Gather 3 rd Party Data Meets Req s Y Underwrite & Make Offer without additional testing Audit N N Initial Screen TrueRisk Life Score Intermediate TRL Scores Apply Full Underwriting High TRL Scores Order Additional Req s Additional Information Needed 14

15 Accelerated Insurance Timeline What s been happening Where are we headed Data Testing -Capturing Requirements -Potential Pilots -Pilots running (dozens) -Early Production -Data Enrichment -12/31/17 Deadlines Production Expansions -Slow Mover Catch-up -Early Insights from Pilots 15

16 The Data 16

17 New Data Sources Data overload Available? Useable? Relevant/Reliable? Consumer (purchases) Credit (FCRA) Medical (blood profiles, histories) Legal boundaries Reputation Responsible use Protective Value Studies Lifestyle Individual/Household Personal (genetics, wearables, social media) Databases Rx, MIB, MVR What s relevant depends on where and why the data is getting used 17

18 Data Considerations Internal Data Industry Management Validation Protective FCRA vs. Value & Non-FCRA Exclusivity Predictive Models The future will be data driven We have only scratched the surface on data solutions Do your homework no all data is created equal 18

19 Industry Validation Internal Data Management Protective Value & Exclusivity Industry Validation FCRA vs. Non-FCRA Predictive Models Retrospective Analysis Demonstrates the value on historical experience mortality/lapse/etc. Provides a starting point in setting future assumptions Should be considered necessary for the validation of a new data source Distribution Analysis Caution when using alone! Anything can segment your business Takes many years to validate experience Combined with a retrospective study can provide great insight 19

20 FCRA vs. non-fcra Internal Data Management Protective Value & Exclusivity Industry Validation FCRA vs. Non-FCRA FCRA Fair Credit Reporting Act (FCRA) Section 604 specifies permissible purposes for use of consumer reports to a person which it has reason to believe intends to use the information in connection with the underwriting of insurance involving the consumer Predictive Models Non-FCRA Generally used in lead generation/target marketing Must understand whether the data is for an individual or household Data quality may not be as accurate as FCRA compliant data 20

21 Predictive Models Internal Data Management Protective Value & Exclusivity Industry Validation FCRA vs. Non-FCRA Understand the development Target variables, input data, modeling technique, etc. More data inputs does not always indicate a better model Is your data source raw data or modeled data? Predictive Models Model should be transparent A score from a model should also provide the drivers behind the score A score from a model should have meaning Transparency is key How will you communicate the results? Ask the tough questions! How much data was used to build and validate? Was the data related to the data that will be used going forward? Did the model really address my problem? Example: Predicting death or underwriting decision? What happens if a data element is not available? Did your modeler understand the business? Do they have a stake in the success of the model? 21

22 Protective Value & Exclusivity Internal Data Management Protective Value & Exclusivity Industry Validation FCRA vs. Non-FCRA Protective Value Determine the value the data provides to you The value should always out way the cost of the data Predictive Models Exclusivity Data may illustrate protective value on its own, but The protective value may diminish with current pieces of data in practice 22

23 Internal Data Management Internal Data Management Protective Value & Exclusivity Industry Validation FCRA vs. Non-FCRA How do we effectively manage and expand our internal data assets to: Properly leverage existing data assets Maximize the value of existing data Do NOT get left behind Leave no stone unturned! Look within your company for ALL potential data sources Be inquisitive Have we captured AND optimized all of our data? Consider investing in data Capture missing historical data Invest in new, emerging data assets Predictive Models Data Management Strategy Develop a corporate strategy to ensure enterprise alignment and minimize duplicative efforts 23

24 Case Study - Credit TransUnion TrueRisk Life 24

25 Big Data Universe Filtered to Get the Most Important Attributes Credit data 800 credit attributes that quantify risk associated with access to credit. 25 credit attributes that quantify behavioral risk associated with credit, foreclosure, bankruptcy and eviction. Marketing Shopper (credit card transactions) Social media Household information Checking/savings IRS data Income, race, ethnicity Public Records Criminal Derogatory records Court filings Data not used in the model Titles and Licenses Property ownership Professional licenses 25

26 Model Creation Building the Model TransUnion & RGA built and tested TrueRisk Life on 92 million individuals Starting Data Variable Selection Model Process External Validation of Model TrueRisk Life Score Built the model on 44 million lives and over 3 million deaths Started with over 800 variables offering features of individual s credit history Selected variables that were: Most predictive of the outcome Stable over time Non-gameable Not too correlated with the other variables Binary Logistic Regression Model validated internally using an additional 30 million lives Tested model using traditional mortality and lapse studies Used a random holdout dataset of another 18 million lives TrueRisk Life presented as a score from: 1 to 100 Low Risk High Risk FCRA Compliant Data Inputs Model calibrated to actual deaths, not underwriting decision Each model score represents 1% of population Achieving 98% scored rates within fully underwritten environments 26

27 TrueRisk Life Deliverables What value is a model if you do not understand the drivers? 1 TrueRisk Life Score 2 Reason Codes 3 Credit Report 27

28 Model Validation Initial Population Study Overall Mortality Details Mortality study performed on holdout sample of 18 million lives using a 1998 TransUnion archive and studying the lives during Study shows 5 times segmentation ( compared to 1-5) Retrospective study provides validation that the model truly predicts mortality 28

29 Model Validation Insured Lives Study Retrospective study on insured lives provides necessary validation Segmentation exists within risk classes; Mortality for worst TRL scores (71-100) are about double that of best risks (1-10); Non-smokers are shown, but results are similar for smokers. Term, UL & VUL; Face Amounts $100,000; Issue Ages < 70 29

30 Model Validation Insured Lives Study Must understand all experience based on the new data Segmentation of about 6 times seen in first two durations within given risk class Non-smokers are shown, but results are similar for smokers Term, UL & VUL; Face Amounts $100,000; Issue Ages < 70 30

31 Model Validation Insured Lives Study Distribution analysis Distribution of Insureds (Compared to Population) Details of the Study Distribution analysis must be done in order to understand how a model or data point can segment business, however; A distribution analysis does NOT provide insights into the future experience of a block without the value of a retrospective study 31

32 Key Takeaways Accelerated Underwriting: Spreading quickly Partnerships expediting the process Experience too soon to tell Data Driven Enabling our new programs Not all data is created equal Invest & Learn 32

33 2015 RGA. All rights reserved.