Casualty Actuarial Society

Casualty Actuarial Society Automated Vehicle Task Force (CAS AVTF) Wenwen Salerno, ACAS, MAAA August, 2015

Vehicle operates as Windows System Automobiles would frequently crash for no apparent reason. This would be so common that motorists would simply accept it, restart their car and continue driving. Airbags would ask, 'Are you sure?' before deployment. Vehicles would occasionally shut down completely and refuse to restart, requiring motorists to reinstall their engine. www.vtpi.org/avip.pdf 2

Example Video 3

CAS AVTF: Overview Goal The CAS AVTF is researching the technology s risks to provide policymakers with the information needed to ensure the product is brought to market as safely and efficiently as possible. Focus Pre market Post market Post claim identify & quantify risks accurately price the technology compensate claimants fairly & efficiently 4

Agenda Automated Vehicles Background Insurance Industry Impact Automated Vehicle Risk Profile Vehicle Symbol Analysis Regulatory Overview 5

Agenda Automated Vehicles Background Insurance Industry Impact Automated Vehicle Risk Profile Vehicle Symbol Analysis Regulatory Overview 6

Automated Vehicle Components Automatic transmissions Diverse sensors (cameras, radars, etc.) Wireless network Navigation system including GPS Automated control (Steering, braking, signals, etc.) Servers, software and power supply www.vtpi.org/avip.pdf 7

Driver Assistance Features 8

Three key technologies for AVs V2V/V2I Vehicle to Vehicle or Vehicle to Infrastructure Dedicated Short Range Communications (DSRC), similar to wifi Vehicle to communication to other vehicles or infrastructure (traffic signals, toll booths, etc LiDAR Light Detection And Ranging combination of light and radar, and uses laser light to create 3D images of the surrounding environment. Remote sensing technology to measure distances Inertial Navigation Systems & GPS INS uses computers, accelerometers (motion), and gyroscopes (rotation) Calculates position, orientation, and velocity 9

Cameras and Deep Learning may be another technology NVIDIA @ CES 2015 NVIDIA video https://www.youtube.com/watch? v=o29tby2a0ek 10

Historic Developments 2005 2010 2011 2012 2013 2014 05: Stanford wins DARPA Grand Challenge 07: CMU wins DARPA Urban Challenge 09: Google begins tescng on public roads 09: EU launches Project SARTRE 10: Volvo City Safe standard Google surpasses 150K miles BMW begins tescng self driving car on public roads NV passes autonomous car law Google @ 300K accident free miles Nissan opens research facility in Silicon Valley Google & ConCnental receive autonomous car licenses FL & CA pass autonomous car laws Google @ 500K miles Oxford creates a $7,750 self- driving system Public road tescng Britain Mercedes CMU Audi receives AV car license NHTSA issues policy on AVs DC passes AV car law MI passes AV law NHTSA passes V2V Google @ 700k miles Volvo Drive Me tests in Gothenburg Google developing driverless car without steering wheel or brakes 11

2015 Developments January Audi partner with Nvidea to use Tegra X1 chips Nissan & NASA announce automated research project partnership February Uber and Carnegie Mellon University announce strategic partnership Apple reported to be working on automated vehicles Sony s image sensors and ZMP s robotics April Delphi 3,400 mile cross country trip with 99% of miles in autonomous mode Nevada licenses Mercedes Freightliner truck Nokia s HERE: 2 billion offer from auto manufacture and Uber reported to offer $3 billion May Tesla adjusts adaptive cruise control to limit liability Daimler s self-driving truck got license and hit the road July M City in University of Michigan 12

London s AV testing regulations may increase investment No special licenses or permits requires No geographical limits No additional insurance requirements Goal Light-touch non-regulatory approach provides clarity for industry to invest in further in research 13

London has 3 trials underway UK Autodrive Programme: 3 years to pave way for introduc?on of AVs Dept. of Transporta.on put ~$29M USD for trials Explore both legal and technical changes required for Autonomous Vehicles Milton Keyes and Coventry Lutz Pods that drive in pedestrian zones Max speed 15 mph Electronic AV Greenwich GATEway shucles Electronic AV Local tour with drop off points: input descnacon on CPU Bristol Venturer consorcum will invescgate congescon and safety BAE Wildcat 14

Future development may create three models for AVs Private: all driving/limited loca?on End to end service Only operates in specified area Taxi service Google, Uber Private: Some driving/all loca?ons Takes over some of the driving E.g. Supercruise, parallel parking Only operates in specified area Driver owns and operates Mercedes, BMW, Volvo, Cadillac, Telsa

Commercial: Some driving/special Vehicle http://www.bloomberg.com/news/articles/2015-05-14/daimler-s-freightliner-tests-self-driving-truck-in-nevada 16

Commercial: Some driving/special Vehicle Daimler s video 17

Agenda Automated Vehicles Background Insurance Industry Impact Automated Vehicle Risk Profile Vehicle Symbol Analysis Regulatory Overview 18

Private Passenger Auto Insurance Coverage Components PRIVATE PASSENGER AUTO INSURANCE EXAMPLE Other - Addi?onal Service, $17 Physical Damage - Collision, $105 Liability - Bodily Injury Liability, $131 Physical Damage - Comprehensive, $50 Liability - Property Damage Liability, $78 Liability - UM/UIM Bodily Injury, $85 Liability - Medical Payment, $7 Liability - Personal Injury Protec?on, $104 From presenter s example quotes 19

Physical Damage/Liability Coverages 20

Physical Damage/Liability Coverages 21

Liability Coverages 22

Private Passenger Auto Liability vs. Products Liability ALAE Factor Permissible Loss Ratio Accident classification Liability limits Settlement lag Unnecessary coverages 23

Insurers responses Pricing Adjustments Proprietary coverage level vehicle symbols Paten?ng Pricing Approaches State Farm: Trip- Based Insurance Pricing Plan (2015) Travelers: Risk- Zone Pricing (2014) Progressive: Vehicle sensor approach (2012) Forming Partnerships Ford, State Farm & U of Michigan Ford Hybrid automated research vehicle (Dec 2013) Honda & major insurance company sign agreement to use self- driving automobile test track at former Concord Naval Weapons StaCon (March 2015) Tes?fying at hearings CA DOI: State Farm & NaConwide & CAS AVTF NJ Senate: Munich Re America Industry Groups are performing research CAS; HLDI- IIHS; RAND Corp; Brookings InsCtute 24

Agenda Automated Vehicles Background Insurance Industry Impact Automated Vehicle Risk Profile Vehicle Symbol Analysis Regulatory Overview 25

93% of accidents are caused by human error. 26

National Motor Vehicle Crash Causation Survey (NMVCCS) Limiting Factors 50% Technology Issues Behavioral (Driver) Issues 48.9% 40% 30% 20% 10% 0% 32.4% 21.3% 16.7% 12.2% 11.6% 11.0% 0.4% 3.1% 2.3% 2.9% 1 2 3 4 1 2 3 4 5 6 27

NMVCCS Implications of the CAS Study New benchmark should be calculated Data is old and unrepresentacve Human driving risks <> automated vehicle risks Appropriate test for each risk Computer simulacons for technology s error rate SimulaCons provide licle insight into driver s actual use of technology. Policy changes can increase AV s safety 1% reduccon in accidents is ~55k fewer accidents and $1.4 billion of economic value per year Policy cost benefit analysis E.g. driver training program, automated vehicle only lanes, allowing the Avs to speed 28

Agenda Automated Vehicles Background Insurance Industry Impact Automated Vehicle Risk Profile Vehicle Symbol Analysis Regulatory Overview 29

Vehicle symbol analysis approach Vehicle experience groups Complements to credibility Each group s experience is weighted and combined with similar vehicles Vehicle s body style factor Prior year factor Automated vehicle symbol: op?on 1 Automated vehicle symbol: op?on 2 Assume a brand new vehicle e.g. Mercedes introduces a new fully automated vehicle No inical prior year factor, growth trend impacts credibility Assume update to a current vehicle e.g. all new Honda Civics sold with AV equipment 30

Vehicle symbol analysis approach Vehicle experience groups Complements to credibility Each group s experience is weighted and combined with similar vehicles Vehicle s body style factor Prior year factor Automated vehicle symbol: op?on 1 Automated vehicle symbol: op?on 2 Assume a brand new vehicle e.g. Mercedes introduces a new fully automated vehicle No inical prior year factor, growth trend impacts credibility Assume update to a current vehicle e.g. all new Honda Civics sold with AV equipment 31

Vehicle Symbol Calculation Automated vehicle symbol: op?on 1 Number of Exposures Assume a brand new vehicle e.g. Mercedes introduces a new fully automated vehicle No inical prior year factor, growth trend impacts credibility Vehicle Symbol Discount Loss Aeenua?on Year 0% 25% 50% 75% 100% 2,500 1 0.0% 0.5% 0.9% 1.3% 1.8% 5,000 2 0.0% 1.4% 2.6% 3.9% 5.1% 7,500 3 0.0% 2.8% 5.1% 7.4% 9.7% 10,000 4 0.0% 4.4% 8.0% 11.6% 15.2% 32

Vehicle Symbol Calculation Automated vehicle symbol: op?on 2 Assume update to a current vehicle e.g. all new Honda Civics sold with AV equipment Vehicle Symbol Discount Loss Aeenua?on Year 0% 25% 50% 75% 100% 1 0.0% 4.3% 7.4% 10.5% 13.6% 2 0.0% 7.1% 13.7% 20.0% 26.3% 3 0.0% 9.7% 18.2% 25.7% 35.4% 4 0.0% 11.1% 21.0% 31.0% 41.2% 33

Upcoming Projects: Average Accident Rate Comparable Driver Match location & type Match driver characteristics Issues with NHTSA Data Only include police reported accidents Cannot segment by driver type Insurance Data Calculate frequencies for different driving segments Can more accurately define good driver GLM s lead to a more stable & accurate calculations 34

Upcoming Projects: Auto Liability vs. Products Liability Quantify the change in costs from liability systems Scenario #1: Assume no change in accidents Scenario #2: Accident frequency is reduced by X% Determine what X needs to equal for Scenario 1 = Scenario 2 Scenario #3: Cap liability to $Z Determine what Z needs to be for Scenario 1 = Scenario 3 Scenario #4: Combination of Scenario #2 & #3 35

Agenda Automated Vehicles Background Insurance Industry Impact Automated Vehicle Risk Profile Vehicle Symbol Analysis Regulatory Overview 36

Current U.S. regulatory approach varies by state Comments CA, DC, FL, MI, NV have regulations that permit operation/testing of AVs May 2013 NHTSA publication Statement with guidance to states on AV regulations Statement outlined NHTSA plans for testing AV technology 37 hcp://cyberlaw.stanford.edu/wiki/index.php/automated_driving:_legislacve_and_regulatory_accon

Consumer Protection NHTSA States establishes regulacons that manufacturers must self- cercfy with State regulators can impose addiconal requirements Individuals Individuals can sue manufacturers if an error occurs Manufactures An alliance of twelve automakers will create a center for sharing informacon and analysis. 38

Regulatory approach needs to be updated Insufficient protec?on for consumers and manufacturers AlliedSignal, Inc v. Moran NHTSA does not have the scale: annual budget $M 828 22 60,000 Auto Insurance NHTSA IIHS-HLDI NHTSA lacks financial incencve of auto insurance industry Accurate evaluacon of risk Claims handling incen?ves CompensaCng claimants fairly and efficiently is a core competence of the insurance industry Auto industry s competence is not insurance 39

Questions and Discussion