Eric Arnold, Sutherland Doug Morrin, Prudential Eric Myers, AIG Consumer Insurance Mary Jane Wilson-Bilik, Sutherland October 13, 2015 Real Life Minority Report? Life Insurers Grapple With Big Data s Implications 2015 ACLI Annual Conference Chicago, IL
Overview Big Data Definitions Concepts and Uses Disrupters Data Aggregation Unfair Discrimination Privacy Data Brokers Internet of Things Security of Data SEC/FINRA Rules Unfair Trade Practices Consumer Laws Price Optimization 2 Regulatory Issues and Concerns Predictive Modeling in Life Insurance Underwriting Marketing Distribution and In-force management
What is Big Data? Aggregate large amounts of consumer data across the enterprise Apply proprietary algorithms to analyze data and to build models to better predict future trends and behavior Types of Data: 3
How is Big Data Being Used? To build and improve customer Relationships ( CRM ), where the customer might be: Distribution firms Institutional partners/vendors Existing and prospective retail customers To improve product pricing, underwriting and claims management: More accurate risk classifications Reduce risk of adverse selection Better detect fraudulent claims 4
Our New Normal -- Headlines Big Data Tool Analyzes Intentions: Cool or Creepy 5
The Big Questions The Big Questions: To what extent will we trade privacy for instant gratification and convenience? What precautions can you take to ensure your algorithms are fair and not improperly discriminatory? Will we so refine our risk classifications that we price insurance products based on a class of one? How do we manage the disruptions Big Data will mean for our business model? 6
Privacy and Consumer Behavior: What is the Price? 7
The Helpful-Creepy Continuum Needs Anticipation Loyalty Programs Liking Suggestion Lists Behavior Modeling Behavioral Advertising Geo-Tracking Cookies 8
The Buzzword: Disruption 9
Drivers and Disrupters Big Data and computational capacity Use of the cloud and Software as a Service (SaaS) Technologists and the wild west of innovation Culture clashes Millennial influence Internet of Things/Wearables Social Media/Facebook -- Liking Amazon -- Next Best Offers Competition from data aggregators Data brokers Google Hackers from foreign nation states 10
Predictive Modeling in the Insurance Industry Due to rapid improvements in computation power, data storage capacity and statistical analysis techniques, over the last several decades predictive modeling has come into widespread use by corporations looking to gain a competitive advantage. Property and Casualty Insurers have used predictive models to support underwriting, marketing and fraud prevention. Life insurers are catching up and are very interested! 11
Life Insurance: Data and Modeling Uses Underwriting Use algorithm to approximate mortality risk Additional underwriting & investigation Quick approval and rating DNA testing and genetic markers Marketing Highest rated customers Most likely purchasers Next best offers Conservation efforts 12
Life Insurance: Data and Modeling Uses Distribution Management Who is the most profitable? What products and incentives work with which cohort of producers? Where is the best use of wholesaling dollars? In-force Management Monitor changing mortality risk Inform reinsurance decisions Focus retention efforts Inform changes in nonguaranteed elements Fraud detection 13
Legal Construct: Big Data Underwriting Data collection and aggregation Fair Credit Reporting Act requirements 14 SEC and state law data sharing limitations Privacy Consent to share data among affiliates State law limitations Unfair Discrimination Discrimination involving either Security protected classes or between individuals of the same class and equal expectation of life How secure are your systems and those of your vendors and their clients/vendors?
Fair Credit Reporting Act (FCRA) FCRA governs the use in insurance underwriting of information supplied by consumer reporting agencies Includes information about a person s credit history, medical condition (MIB), driving record, criminal activity Consent requirements Consumer rights to correct data and know specific data that resulted in paying more or receiving less coverage Use restrictions: Medical information from consumer reporting agencies used in underwriting cannot be used for marketing Ongoing duty to correct data provided by consumer reporting agency if inaccurate or incomplete 15
FCRA Gives Rights to Consumers FCRA prohibits the sharing of medical consumer report data across affiliates in some circumstances Consumers must be able to opt-out of affiliate sharing of other consumer report information used for marketing purposes SEC Reg. S-AM extends this rule to broker-dealers and investment advisers Can use constructive sharing arrangements FCRA does not apply to most first parties that collect consumer information Does not cover IoT devices Does not cover the selling of anonymous data Adds to complexities of structuring data warehouses 16
Regulators Concerns about Data Brokers Data Brokers: A Call for Transparency and Accountability, Federal Trade Commission (5/14) Lack of Transparency: Data brokers collect consumer data, largely without the consumer s knowledge Complexity: Data broker industry is complex, with multiple layers of data brokers providing data to each other Vast Scope: Data brokers collect and store billions of data elements on early every U.S. household and commercial transaction Make Inferences about Consumers, Some Sensitive: Data brokers infer consumer interests and characteristics based on data they collect: Expectant Parent, Diabetes Interest Inaccessible: If errors, consumers can t correct the problem 17
Internet of Things (IoT) Consumer-facing devices that collect wearables data Wearables data is data from devices worn on the body, connected through apps on a mobile device, and can include data around the wearer and about the wearer. Bracelets that share with your friends how far you biked or ran that day Connected medical devices that allow consumers with health conditions to connect to their physicians about their disease Aggregated by data brokers and sold 18
Regulators Concerns with IoT Internet of Things: Privacy & Security in a Connected World, Federal Trade Commission Staff Report (1/15) Cybersecurity Risks: unauthorized access and misuse of personal information and safety risks Recommend: Vendors must patch known vulnerabilities over life-cycle of product Privacy Risks: sensitive personal data and behavioral patterns might be collected Concern that it could used to make credit, insurance and employment decisions without the customer s knowledge Data minimization: FTC staff -- principle of privacy protection Notice and Choice 19
Privacy Laws Privacy laws require sending privacy notices informing customers of how the company will use their personal identifiable information (PII) Under some laws, customers must be given an opt-out right to the sharing of PII Some state laws are more restrictive Legal, contractual (terms of use), and reputational issues even where consumers have no legal right of privacy Insurers have tended to silo data. If the insurer want to aggregate PII across affiliates and/or cross-sell: What do your applicable privacy policies say? What do your selling agreements say? 20
Anti-Discrimination Laws Unfair trade practices laws prohibit unfair discrimination between individuals of the same class and equal life expectancy when underwriting life insurance Suspect classification laws provide that using factors such as race, color, religion, national origin or gender in the underwriting process will also be viewed as unfairly discriminatory even if actuarially justified. Concern: How do you prevent algorithms from including proxies for information that cannot be used in insurance underwriting? Who is watching the technologists for social bias? 21
Same Class (Actuarial) Discrimination A fundamental principle of insurance underwriting, pricing, reserving is that like risks should be treated alike. How to handle a Pool of One? Prior to recent state bulletins on price optimization, insurance departments had made no mention of unfair discrimination in the context of predictive analytics 22
Price Optimization in the Crosshairs Price elasticity and how much will you pay? Regulators: California; District of Columbia; Florida; Maryland; Indiana; Maine; New York; Ohio; Pennsylvania; Rhode Island; Vermont Focus to date solely on Property/Casualty National Association of Insurance Commissioners White Paper on Price Optimization Class Actions 23
Digital Distribution and Sales Helpful v. Creepy Continuum SEC/FINRA/State insurance advertising rules Consumer Protection Statutes.com Disclosures (FTC 2013 Report) CAN-SPAM Relationships with SaaS Providers (Salesforce) and Social Media Sites Relationships with current distribution partners 24
Big Data Questions Before forging ahead, answer these questions: Where are we getting the data? Is the source of the data authorized to provide it to us? Is the data reliable? Are we authorized to use it? What do we want to use the data for? Is it personalized or de-identified/anonymized? Who is checking the algorithms for social bias? Do our policies communicated to our customers and business partners allow us to use the data? How will our customers react when they understand, or can guess, what we did? How will we protect the data? 25
Questions? Mary Jane Wilson-Bilik Sutherland 202.383.0660 mj.wilson-bilik@sutherland.com Eric Myers AIG Consumer Insurance 713.831.6225 eric.myers@aig.com Eric Arnold Sutherland 202.383.0741 eric.arnold@sutherland.com Doug Morrin Prudential 973.802.6025 douglas.morrin@prudential.com 26