Accelerated Underwriting. Murali Niverthi, PhD, FSA, MAAA Assistant Actuary, Integrated Underwriting Solutions

Transcription

1 Accelerated Underwriting Murali Niverthi, PhD, FSA, MAAA Assistant Actuary, Integrated Underwriting Solutions

2 Agenda 1 2 Mortality Current landscape considerations and preliminary findings 3 The years ahead June 21,

3 Current landscape 1 Image licensed from Getty Images

4 History of automated underwriting: Confluence of a number of factors Customer experience Underserved middle market Faster, less intrusive Expectations from Millennials Big data Public databases with personal, credit and financial information Computing power and predictive analytics Insurer and producer Faster, cheaper, more sales Less administrative work June 21,

5 Evolution of the new underwriting paradigm Industry-wide Today Programs Few mostly SI Handful some accelerated Most varied Underwriting tools MIB, MVR, Rx Early models, interviews, reflexives More models & electronic health records Rules engines Some Half Most Pricing Table % loads Fully-underwritten Face amounts Up to $100K $250K $1M+ June 21,

6 Accelerated underwriting process Insurance History Prescriptions Driving Record Credit Electronic health records Lifestyle / Social Wearables Application & Tele- Interview Predictive models: 1) screen 2) risk selection 3) smoker Rules-based Automated UW Holdout Manual UW Risk Class Traditional Data/Process Existing Data Sources New Data Sources New Process Medical Lab Results Attending Physician Statement Income & financial info June 21,

7 Mortality considerations and preliminary findings 2 Used by permission

8 Factors affecting AUW mortality Databases Application Target market/ Distribution channel Fully UW mortality Smoker veracity No fluids Antiselection / sentinel effect Accelerated underwriting mortality June 21,

9 Impact of no fluid testing, anti-selection and smoker veracity Medical Underwriting Accelerated Underwriting Tobacco use self-reported & verified self-reported Personal history self-reported & verified self-reported with limited verification Blood pressure Verified self-reported Blood chemistry Total chol / chol ratio Verified self-reported or N/A Liver enzymes Verified N/A HgA1C Verified N/A Drug/alcohol abuse self-reported & verified self-reported MVR DUI self-reported & verified self-reported & verified Moving violations self-reported & verified self-reported & verified Family history self-reported self-reported Build Verified self-reported June 21,

10 Tobacco misrepresentation Mortality impact of undisclosed tobacco use NT in insured block 90% TB misrepresentation 30% TB/NT mortality load 200% Additional load to be applied to NT risk class to account for extra mortality of undisclosed TB users: > 3% June 21,

11 Target market and distribution channels Credit-based mortality scores BMI distribution skewed Tobacco user proportions Financial underwriting needed Unemployed Blue collar/unskilled labor Face amounts Image licensed from Getty Images June 21,

12 AUW approval rates by pricing cell Final Class AUW Eligible AUW Approved Fully UW (with RHO) Total Super Pfd 42% 45% 55% 100% Preferred NT 52% 38% 62% 100% Standard NT Preferred Tb Standard Tb Total 41% 35% 18% 82% 100% Gender Male 38% Female 48% 61% 46% 54% 100% 26% 32% 68% 100% Account for age, face amount, etc Total % 44% 18% 9% 36% % 46% 29% 14% 36% % 48% 18% 12% 31% 1 million 47% 30% 14% 6% 22% June 21,

13 Random holdout (RHO) sample June 21,

14 Early findings: summary Metric Range Trend Future?? Mortality Nothing credible TBD Looks almost identical to pre AUW AUW approval rates % Mid 30s 50%+ by 2020 Lapse Improved placement 0-4% RHO (tobacco misrep among RHOs determined to be smoker) RHO (misclassification best preferred) Higher than FUW 25 55% Concentrated in early durations. Not credible Same and watch distribution and customer expectations 1-2% smoker approvals vs. overall Trend toward FUW, then improved through personalization and engagement Improved through personalization Smoker prediction 15 30% Bulk off 1 2 classes Materially better accuracy June 21,

15 Validation of new risk assessment tool standard Assign cost to misclassification preferred June 21,

16 Sources of kick-out to full underwriting Part I Part II MIB MVR Rx check Credit based score Predictive model June 21,

17 The years ahead 3 Licensed from Getty Images

18 The years ahead What changes are coming? Increasingly advanced predictive models and rules engines Dynamic monitoring processes preand post-issue Face amount ranges, risk classes Niche target markets Risk mitigation challenges obesity, smoker misclassification, opioid abuse, proliferation of genetic testing. New data sources Distribution channels Be prepared to make adjustments June 21,

19 The years ahead The rise of the engaged consumer Targeted, scored, educated Temporary and portable Ongoing refinement Lifecycle view Underwritten and engaged Multiple touchpoints Continually underwritten Health and lifestyle portal June 21,

20 Summary Current landscape Mortality considerations and preliminary findings The years ahead June 21, 2018 Image licensed from Getty Images