An Analysis of Economically Efficient Insurance Schemes for Automated Vehicles

  An  Analysis  of  Economically  Efficient  Insurance  Schemes   for  Automated  Vehicles               Brandon Xavier Rhodes June 2014 Advised by Professor Alain Kornhauser Submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Engineering Department of Operations Research and Financial Engineering Princeton University       I hereby declare that I am the sole author of this thesis I authorize Princeton University to lend this thesis to other institutions or individuals for the purpose of scholarly research Brandon Xavier Rhodes I further authorize Princeton University to reproduce this thesis by photocopying or by other means, in total or in part, at the request of other institutions or individuals for the purpose of scholarly research Brandon Xavier Rhodes   -­‐  ii  -­‐   Acknowledgements   I would like to first thank my family for being there to support me throughout my Princeton journey and this process in particular Your guidance has been invaluable I would also like to thank Allie, whose gestures of support, from cupcakes to words of encouragement, provided me with persistent motivation Finally, I would like to thank my advisor Professor Alain Kornhauser who inspired my thesis topic In addition, Professor Kornhauser provided much appreciated guidance throughout my writing and was always available to discuss ideas or concerns I had Professor Kornhauser, your excitement and passion for automated vehicles is evident both in the discussions we had about my thesis and in your lectures It’s been a pleasure to work with you on this research   -­‐  iii  -­‐   Abstract Automated vehicles have the potential to have a profound impact on the automobile insurance industry Preliminary research suggests that various automation systems are correlated with a drastic reduction in the frequency of insurance claims However, American companies have yet to offer unique plans for customers who own vehicles equipped with an automation system, thereby resulting in economic inefficiency This thesis suggests various insurance schemes that are designed to insure automated vehicles in an economically efficient manner First, this thesis provides an overview of the automobile insurance industry and an overview of automated vehicles Afterwards, the dilemma automated vehicles pose to current liability laws is also examined This is because the legal treatment of vehicles equipped with automation systems will have a significant effect on how they are insured Next, before offering suggested insurance schemes, the results of a preliminary analysis on the effect of automation systems on insurance claims are examined The thesis concludes with a theoretical analysis of the impact automation systems will have on various sectors of the economy once they achieve greater market penetration   -­‐  iv  -­‐   TABLE OF CONTENTS Introduction Chapter 1: Current Automobile Insurance Schemes 1.1  Basic  Monthly  Premium     1.2  Pay-­‐As-­‐You-­‐Drive   Chapter 2: Probability Models 13 2.1  Claim  Frequency   2.2  Claim  Severity     Chapter 3: The Legal Dilemma Posed By Automated Vehicles 17                                                                                             3.1  Common  Carrier  Laws  Applied  to  Automated  Vehicles       3.1.1  Implications  for  the  Insurance  Industry     3.1.2  Implications  for  the  Market   3.2  Product  Liability  Laws  Applied  to  Automated  Vehicles       3.2.1  Implications  for  the  Insurance  Industry     3.2.2  Implications  for  the  Market   3.3  Risk-­‐Utility  Test  Applied  to  Automated  Vehicles       3.1.1  Implications  for  the  Insurance  Industry     3.3.2  Implications  for  the  Market   3.4  Cost  of  Litigation     3.5  Conclusion     Chapter 4: HLDI Research Study Of Crash Avoidance Systems 28 4.1  Acura     4.2  Buick   4.3  Mazda     4.4  Mercedes-­‐Benz   4.5  Volvo   4.6  Volvo  City  Safety     Chapter 5: HLDI Research Study Results categorized by Crash Avoidance System 65 5.1  Blind  Spot  Assistance  Systems     5.2  Curve  Illumination  Systems     5.3  Lane  Departure  Warning  Systems     5.4  Forward  Collision  Warning  Systems  without  Automated  Braking     5.5  Forward  Collision  Warnings  Systems  with  Automated  Braking     5.6  Parking  Assistance  Systems     5.7  Conclusion     Chapter 6: Examination of the Availability of Forward Collision Warning System 80 6.1  Availability  to  Consumers     6.2  Manufacturer’s  Commitment  to  Offering  the  Technology       -­‐  v  -­‐   Chapter 7: Suggested Insurance Schemes for Vehicles Equipped with Automation Technology 97 7.1  Level  0  Automation  Technology     7.2  Level  1  Automation  Technology     7.3  Level  2  Automation  Technology     7.4  Level  3  Automation  Technology     7.5  Level  4  Automation  Technology     Chapter 8: Financial Implications of Passenger Vehicles Equipped with Automation Technology 115 Chapter 9: Conclusion 120 9.1  Limitations     9.1.1  Technology     9.1.2  Legal     9.2  Final  Conclusions     9.3  Suggestions  For  Further  Research     Chapter 10: Literary Review 131 Bibliography………………………………………………………… ………… … 135           -­‐  vi  -­‐     Introduction The notion of cars driving themselves is not a new idea For sometime, people have fantasized about cars that could drive without human input The advent of autonomous cars would not only revolutionize the way people travel, but would also revolutionize many industries, resulting in a drastic impact on the global economy Some of the industries that would be affected include, but are not limited to, car manufacturing, transportation services, and the automotive insurance sector This paper will focus on the automotive insurance industry Specifically, it will focus on how to price insurance for autonomous cars Before one considers how to price insurance plans for autonomous cars it is important to first understand the current progress of the technology and the legal implications of introducing cars that drive without human input It is also beneficial to consider how insurance companies currently price insurance plans, because aspects of the current pricing method can be used when pricing plans for autonomous cars Before continuing, it is important to realize that there are varying levels of automation Thus, the National Highway Traffic Safety Administration (NHTSA) has released a report detailing what defines these levels The first level is At this level, according to NHTSA (2013), drivers are in complete control of the vehicle and there is no automation (p.5) At level 1, NHTSA (2013) states that the vehicle has “functionspecific automation” (p.5) In the report, NHTSA mentions some examples of functions at this level such as stability control, cruise control, automatic braking, and lane keeping (p.5) At level 1, the automation technology is classified as driver assistance technology   -­‐  1  -­‐   Therefore, the driver is to be in control of the vehicle and maintain full attention while driving NHTSA asserts that at this level, the driver is not to have both hands off of the wheel and feet off of the pedals at the same time (NHTSA, 2013, p 5) Level automation “involves automation of at least two primary control functions designed to work in unison to relieve the driver of control of those functions.” (NHTSA, 2013, p 5) Thus, at level 2, the driver is able to have both hands off the wheel, feet off of the pedals, and cede control of primary driving functions to the vehicle (NHTSA, 2013, p.5) However, the driver is to maintain alertness and must be ready to take immediate control of the vehicle if necessary (NHTSA, 2013, p.5) Level automation is characterized by vehicles with the capability to “enable the driver to cede full control of all safety-critical functions under certain traffic or environmental conditions and in those conditions, to rely heavily on the vehicle to monitor for changes in those conditions requiring transition back to the driver.” (NHTSA, 2013, p.5) As in the case of level 2, the driver is to be available for taking control of the vehicle, but a key difference is that the driver has to have a “sufficiently comfortable transition time” (NHTSA, 2013, p 5) To achieve this transition time, the vehicle alerts the driver when it is no longer capable of using automation to drive As NHTSA states, the driver is not expected to monitor the roadway with the same diligence when operating a vehicle with level automation as he or she is expected to use when operating a vehicle with level automation (NHTSA, 2013, p.5) Level is characterized by full automation (NHTSA, 2013, p.5) The vehicle is expected to carry out all aspects of driving for an entire trip whether a driver is present in the vehicle or not (NHTSA, 2013, p 5) A vehicle system with level automation technology is treated as a system that replaces a human driver A vehicle with level   -­‐  2  -­‐   automation is not expected to drive in circumstances that are not suitable for human drivers, such as severe weather conditions The goal of autonomous cars is to have these cars drive as if there was human input from the best possible human driver Therefore, the car would minimize the possibility of having an accident in all situations and would maximize other aspects of driving, such as fuel efficiency In order to so, the cars must be able to perceive the world around them as humans Thus, the cars will need to be aware of their surroundings, the road conditions, traffic lights, lane markings, and so on To achieve this objective, autonomous cars use a combination of sensors that include cameras, radar, GPS, and LIDAR to gather the necessary information from the world around them and to use that information to “see” what is needed to drive in a safe manner The information collected from the sensors is then sent to computers on board the car to be analyzed by algorithms written to understand the information Once the information is analyzed, commands are sent to the systems that control the steering, braking, and acceleration of the car so it can operate in the proper manner For example, a camera mounted onboard the car captures an image of the lane markings on either side of the car, while radar detects an object moving in front of the car The algorithms recognize the image captured by the camera as lane markings, as another algorithm determines that the object moving in front of the car can be classified as another vehicle As a result, commands are sent to the system controlling the steering so that the steering wheel moves in a manner that allows the car to remain in the lane Also, a command is sent to the systems controlling the acceleration of the car, directing it to accelerate enough to keep a safe following distance from the car in front of it This example is clearly a simplification of the process   -­‐  3  -­‐   involved However, it is designed to capture the main idea of one of the key capabilities needed for the autonomous car to function properly The process of interpreting the information collected by the sensors is very technical in nature and as such, is not the focus of this paper The success of autonomous cars is only possible through seamless connections among the hardware sensors, the data they collect, and the sophisticated algorithms needed to process that data, determine the appropriate action, and respond properly The unique hardware necessary for autonomous cars includes sensors to collect information about the surrounding world, along with systems to control basic driving actions, such as acceleration, braking, and steering Currently, the sensors being used by researchers in this field are a combination of radar, GPS, and LIDAR Different companies use different combinations of these sensors, and some companies have attempted to forgo using LIDAR The algorithms that are used to process the information can be grouped into four main categories These categories are: lane detection, object detection, road analysis, and control of systems involved in basic driving actions The details of the algorithms and the mathematics involved are beyond the scope of this paper Consequently, the paper makes the assumption that the algorithms used by the different car companies to program for automation are uniform in their functionality and reliability This assumption aids in reducing the complexity of pricing the insurance plans, for without this assumption one would need to take into consideration the different limitations of each algorithm used, which may affect the performance of the automated function At the time this thesis was written, car companies and other players in the field have made significant progress in developing reliable level automated cars The rapid   -­‐  4  -­‐   9.2 Final Conclusions At present, automated vehicles are offered to consumers and are in use today Currently, the levels of automation used in public vehicles are levels and 1, and to a small extent, level As such, the automation serves to assist the driver in two ways The first is as a convenience feature; the second is as a system in place to aid in the prevention of accidents As mentioned earlier, the technology for fully automated vehicles is currently being tested and it appears that it will be ready for the public marketplace in the near future However, there are obstacles to its introduction The obstacles to fully automated vehicles are financial, legal, and psychological in nature If these obstacles are to be overcome, proper insurance plans for automated vehicles need to be in place to support its adoption in the marketplace The current legal system is an obstacle for the adoption of levels and technology because the current liability laws are not equipped to handle cases of automated vehicles causing an accident Levels through not cause the same legal conundrum as the higher levels because these levels are considered driver assistance technology Therefore, it is presumed that the driver maintains attention to the surroundings and control of the vehicle at all times and is not to depend upon the technology in preventing an accident As such, the driver is liable for accident if he or she is at fault even if the technology does not function as predicted Let’s examine the case of a malfunctioning blind spot assist warning to illustrate this point A driver’s car is equipped with blind spot assist so that when switching lanes a warning activates when another vehicle is in the driver’s blind spot to prevent an accident A driver still has the duty to manually check his or her blind spot before   -­‐  123  -­‐   switching lanes, but it is easy to imagine a driver becoming accustomed to the technology to the point that he or she no longer bothers to check One day, when switching lanes, the blind spot assist fails to activate when another vehicle is present in the blind spot The driver does not manually check for another vehicle and causes an accident Legally, the driver is fully liable for the ensuing accident because he or she failed to manually check the blind spot This example illustrates that levels and automation are not meant to be replacements for drivers or for good driving habits They are meant to assist the driver; since the technology plays a subservient role in that the driver is liable for an accident even if the technology malfunctions The insurance plans detailed in prior sections for driver assistance systems are designed to capture the added benefits they provide to both the driver and to insurance companies Providing discounts to drivers that own vehicles with these systems will increase the rate of market saturation of vehicles with the technology This is advantageous to both insurance companies and drivers because the technology reduces the number of expected accidents for the driver This reduces the expected cost to repair or replace the car for the insurance company, while for drivers it reduces the expected increase in premiums incurred by causing an accident The higher price of a car with a crash avoidance or driver assistance system acts as a barrier, preventing new car purchasers from choosing it versus a similar car without the added feature A discount on the insurance premium mitigates the higher initial purchase cost by allowing for savings on the future cost of ownership, thereby increasing the net present value of the car with crash avoidance systems The discount must be such that the net present value of purchasing a car with a crash avoidance system is greater than or equal to purchasing a   -­‐  124  -­‐   similar car without the technology If this is true, more customers will choose the car with the added safety The reduction in expected yearly costs will increase the value of the insurance company and will raise shareholder value This is true if the cost of the discounts is less than or equal to the expected cost of accidents avoided by the technology The calculations done in previous sections suggest that the cost of premium discounts is indeed less than the expected cost of an accident To conclude, the impact of levels and automation on drivers and insurance companies is unclear because of the uncertainty in how liability will be treated in the legal system Also, there is now the added obstacle of the potential psychological aversion consumers may have to vehicles taking control over driving when they have been accustomed to driving themselves or having another human drive What follows from the introduction of vehicles with levels and automation is strongly dependent on how the legal system adjusts to handle liability of an accident caused by a malfunction of a vehicle using the technology There are two possible scenarios, the first being that driver will be held liable, and the second being that manufacturer or supplier of the technology will be responsible If the legal system chooses to treat the driver of the vehicle that causes an accident due to the malfunctioning level or automation technology as liable, then it will be likely that the public will be slow to adopt the technology This is because there is reduced incentive to purchase the vehicle if an accident caused by the vehicle is determined to be the fault of the driver for purchasing the vehicle Most new car purchasers would rather save the additional money for the technology than purchase it at what is perceived to be an additional risk Insurance companies should offer discounts on   -­‐  125  -­‐   the premiums charged; however, the reduced insurance payments will most likely not be enough to combat the psychological barrier for most consumers It is generally understood that humans prefer to be in control of any given situation Purchasing a car that is fully automated removes the control from the driver and gives it to the vehicle This loss of control could be unnerving for the common car owner due to the fact that the vehicle’s control over one’s safety requires the driver to trust the safety of the technology Even if the trust is there, the driver then needs to decide if the driver trusts the vehicle’s technology to transport him or her safely, more than he or she trusts him or herself Of course, in the case of a drowsy or drunk driver, one can assume that the driver will elect to have the vehicle drive But, in the case of everyday driving, it remains unclear if the majority of people would elect to have the vehicle drive In depth analysis on the psychological barriers that could prevent most consumers from purchasing vehicles with levels and automation is beyond the scope of this paper But, it is recommended that studies conducted to more accurately assess this potential barrier to the adoption of the technology However, even under this scenario, it can be concluded that the car manufacturers have a financial incentive to produce and sell vehicles with this technology and that insurance companies have a financial incentive to have their customers purchase vehicles with this technology Moreover, it can be concluded that if the data is as it is suggested, new car purchasers would financially benefit from purchasing cars with this technology, although it is likely that the rate of adoption will be slow and many consumers will be priced out of the market The uncertainty lies in whether the financial incentive for consumers is large enough to overcome potential psychological barriers   -­‐  126  -­‐   If, however, the legal system chooses to treat the car manufacturer, or supplier of the automation technology, as liable for an accident that is caused by a malfunction then the financial burden shifts from the consumer to the car manufacturer Car manufacturers will have to the net present value calculation to determine if expected earnings of selling cars with this technology is greater than the expected court costs of accidents caused by malfunctioning technology Under this scenario, new car purchasers have a high incentive to not only purchase cars with this technology, but also to have the vehicle control the driving as much as possible, for if the vehicle is driving, the owner of the car is free from liability for any potential accident This is a large financial incentive because the driver would effectively no longer pay car insurance premiums, save for potentially a small premium to cover the instances when the owner is in control Any psychological barriers that may exist on the part of the consumer would most likely be more readily overcome due to the large financial incentive presented under this scenario Insurance companies, in circumstances in which car manufacturers or suppliers of the automated technology are liable for accidents caused by malfunctioning technology, have the same financial incentives as if the owner of the car is deemed to be liable However, the structure of car insurance would most likely have to adapt There are two possible changes that would occur The first is that the car manufacturers would selfinsure One way car manufacturers could achieve this is by setting aside a trust for potential accident victims equal to the expected cost of accidents caused by their malfunctioning automation technology The second possibility is that some insurance companies would begin insuring car manufacturers in addition to the typical car owner The insurance rates offered would be based on the reliability of the hardware used and   -­‐  127  -­‐   analysis of the implemented code’s expected reliability Other factors such as the company’s brand reputation for quality and the location where the vehicle is used may also be considered The conclusions with regard to vehicles with levels and automation technology serve as a basis for further research More in depth analysis should be done once the legal system decides how to handle the issue of liability, after vehicles with the technology are introduced to the public Currently, only select companies are testing vehicles with technology with levels and automation As mentioned earlier, both Google and Audi have been granted legal permission to test their vehicles on public roadways in select states However, the tests are done in optimal conditions and are extremely controlled More data needs to be available before an accurate assessment of the impact of vehicles with levels and automation can be made   9.3 Suggestions for Further Research First, it is suggested that a psychology study is done whose objective is to determine if people trust automated vehicles to operate driving function In speaking with various students, most have expressed apprehension in allowing a vehicle to operate the primary driving functions Future studies thus should examine this apprehension to predict what percentage of the population feels similarly In addition, specific scenarios where people feel comfortable allowing the vehicle to be in control should be examined Also, various methods to alleviate this apprehension should be evaluated If widespread, this has the capability of negatively impacting the rate of market saturation and will thus influence the technology’s financial impact   -­‐  128  -­‐   Research should also be done on the psychological impact of removing the option to disengage level automation technology Once the technology becomes refined, it is in neither the driver’s nor the insurance company’s best interest to allow the driver to disengage the technology This is especially true of automated emergency braking The technology should at least be engaged upon every vehicle start up to force the driver to make the conscious decision to opt out of the technology However, the influence such a decision would have on sales of vehicles equipped with this technology remains uncertain Thus, future studies should be conducted on this matter In addition, further research on the expected financial impact of the introduction of the technology on sectors of the economy is advised Some of the sectors that could be impacted are automobile manufacturers, hospitals, automotive repair and service shops, the taxi and personal driver industry, the insurance industry, and public transportation This thesis is limited in the scope and depth with which it can examine these economic sectors More robust research and analysis is suggested in order to predict with accuracy the impact of levels and automation technology This research could be used to aid the impacted sectors and prepare them to make the necessary adjustments once the technology is introduced This suggested research would also provide additional insight for policymakers deciding whether to foster an environment conducive to the technology’s development This thesis did not analyze the technical aspects of automation in detail Research should be done on the efficacy of different methods used to achieve automated driving This includes examining different hardware configurations and examining different algorithms Insurance companies could use this research to offer discounts depending on   -­‐  129  -­‐   the efficacy of the automation method and algorithms used on the vehicle model to be insured NHTSA could also use this research to regulate which algorithms should be used by automobile manufacturers in a facet of automation if one algorithm is found to be clearly superior Finally, it is suggested that further research is done on modeling insurance for vehicles equipped level and automation technology once the technology is fully developed and more data is available The suggestions posed in this thesis are based upon predictions of how the technology will develop After the technology has been introduced to the public and sufficient data is acquired, a more accurate analysis both on how to model insurance for these vehicles and on how to assess their financial impact on various sectors of the economy can be conducted   -­‐  130  -­‐   Chapter 10   Literary Review NHTSA “Preliminary Statement of Policy Concerning Automated Vehicles” This source provided a detailed outline and description of the different levels of automated vehicles This information was primarily used in the introduction In addition, NHTSA’s description of the levels of automation was used as the basis for subsequent chapters on insurance models, and to some extent in the legal chapters Cyrus Pinto “How Autonomous Vehicle Policy in California and Nevada Addresses Technological and Non-Technological Liabilities” This source was use in the section on the limitations of the technology It provided the insight that there are certain situations that technology has yet to replace human judgment It gave the specific example of choosing whom to place in harm, a child or animal running out into the street or the occupants of the vehicle It also posed the question of how the vehicle should respond once a bug is discovered Finally, this source offered the possibility of terrorist attacks on automated vehicles Dylan LeValley “Autonomous Vehicle Liability- Application of Common Carrier Liability”   -­‐  131  -­‐   This source was used in the legal section of the thesis The work begins with background information on the technology that has been found in other sources It provided the notion of treating manufacturers of autonomous vehicle technology as common carriers In addition to defining the implications of treating manufacturers as common carriers, it is also provided background information on product liability It also provided the insight to think of automated vehicles as an elevator Alexander P Herd “R2DFord: Autonomous Vehicles and the Legal Implications of Varying Liability Structures This source was used in the legal section of the thesis Its main use was providing the insight to consider using the risk-utility test to determine liability in the case of an automated vehicle causing an accident When discussing the risk-utility method the author provides details on what factors are used to evaluate the test and how the test is ultimately executed Jeffery K Gurney “Sue My Car Not Me: Products Liability and Accidents Involving Autonomous Vehicles” This source was used in the legal section of the thesis It provided ample background on product liability laws and how its application may be applied to autonomous vehicles Not having a legal background it was beneficial in reading the description of what product liability law entails It provided a similar but more in depth look at defining manufacturing and design defects compared to LeValley’s work The specific insights used were the cost prohibitive nature of filing a suit against an   -­‐  132  -­‐   autonomous vehicle manufacturer, and the potential for customers to be hesitant when purchasing a car that he or she would be liable for if it caused an accident Todd Litman “Pay-As-You-Drive Vehicle Insurance Implementation, Benefits and Costs” This source was used in the Current Insurance Schemes section of the thesis It provided an overview of the structure of PAYD insurance, how various forms can be implemented, and the costs and benefits or each form Only per mile and GPS pricing methods were used in the thesis It also gave examples of companies that offer PAYD insurance to its customers Judy Feldman Anderson, FSA and Robert L Brown, FSA “Risk and Insurance” This source provided a general overview of insurance and measuring risk Information from this source was used in the Current Insurance Schemes section and in the Mathematical Models section Muhammad Usman Iqbal and Samsung Lim “A Privacy Preserving GPS Pay-As-YouDrive Insurance Scheme” This source was used in the Current Insurance Schemes section It provided insight into the notion of incentive based PAYD insurance plans It also offered a method to eliminate privacy concerns for using GPS to calculate PAYD premiums HLDI   -­‐  133  -­‐   HLDI conducted the research on the effects of crash avoidance systems on insurance loss claims used in this thesis HLDI was one of the first, if not the first, research institutions to conduct an analysis on level and level automation systems Also, at the time of conducting research HLDI was the only source for data in this field of study Before HLDI’s analysis arguments of the effects level and level automation systems would have on insurance claims were speculative in nature The data provided by HLDI are the cornerstone for the arguments made in this thesis As such, pertinent results were aggregated and reproduced in the thesis to provide a reference from which observations and conclusions could be drawn from IIHS IIHS provided the database of the availability of forward collision warning systems on vehicles offered to the American market On the website one is able to select the manufacturer and the year After the selections are made a list is populated that depicts the models offered by the manufacturer during the query year and for each it model it shows whether a forward collision warning system was an installable option The availability of other crash avoidance systems was also shown, but this thesis focused on forward collision warning systems with and without automated braking functionality The database also depicts whether the crash avoidance system was standard on the model, or available as an option that could be purchased The distinction between standard feature and installable option was not made in this thesis Instead, they were combined to convey total availability       -­‐  134  -­‐     Bibliography   Anderson, J F., & Brown, R L (2005) Risk and Insurance Education and Examination Committee of the Society of Actuaries , 1-16   Bordoff, J E., & Noel, P J (2008) Pay-As-You-Drive Auto Insurance: A Simple Way to Reduce Driving-Related Harms and Increase Equity The Hamilton Project (pp 1-48) The Brookings Institution   Boucher,  J.-­‐P.,  Denuit,  M.,  &  Guillen,  M  Models  of  Insurance  Claim  Counts  with  Time   Dependence  Based  on  Generalization  of  Poisson  and  Negative  Binomial   Distributions  Casualty  Actuarial  Society  ,  2  (1)     Cummins, J D (1991) Statistical and Financial Models of Insurance Pricing and the Insurance Firm The Journal of Risk and Insurance , 58 (2), 261-302   Dionne, G., & Vanasse, C A Generalization of Automobile Insurance Rating Models: The Negative Binomial Distribution with A Regression Component Astin Bulletin , 19 (2)   Fatal Analysis Reporting System Encyclopedia (n.d.) Retrieved June 2013, from National Highway Traffic Safety Administration: http://wwwfars.nhtsa.dot.gov/Main/index.aspx   Gourieroux, C., & Jasiak, J (2004) Heterogeneous INAR(1) model with applications to car insurance Insurance:Mathematics and Economics , 34, 177-192   Grabowski, H., Viscusi, W K., & Evans, W N (1969) Price and Availability Tradeoffs of Automobile Insurance Regulation The Journal of Risk and Insurance , 56 (2), 275299   Gurney, J K (2013, October) Sue My Car Not Me: Products Liability and Accidents Involving Autonomous Vehicles Selected Works   Herd, A P (2013, April) R2DFord: Autonomous Vehicles and the Legal Implications of Varying Liability Structures Selected Works   Highway Loss Data Institute (2011) Acura collision avoidance features: initial results HLDI Bulletin , 28 (21)   Highway Loss Data Institute (2011) Mazda collision avoidance features: initial results HLDI Bulletin , 28 (13)   -­‐  135  -­‐     Highway Loss Data Institute (2012) Mercedes-Benz collision avoidance features: initial results HLDI Bulletin , 29 (7)   Highway Loss Data Institute (2012) Volvo City Safety loss experience-an update HLDI Bulletin , 29 (23)   Highway Loss Data Institute (2012) Volvo collision avoidance features: intitial results HLDI Bulletin , 29 (5)   Highway Loss Data Institute (2011) Buick collision avoidance features: intital results HLDI Bulletin , 28 (22)   Howard, B (2013, May 3) Are we there yet: Cadillac's semi self-driving car with Super Cruise Retrieved April 2014, from Extreme Tech : http://www.extremetech.com/extreme/154555-are-we-there-yet-cadillacs-semi-selfdriving-car-with-super-cruise   Insurance Institute for Highway Safety (1996-2014) Crash avoidance features by make and model Retrieved 2014, from http://www.iihs.org/iihs/ratings/crash-avoidancefeatures   Ippolito, R A (1979) The Effects of Price Regulation in the Automobile Insurance Industry Journal of Law and Economics , 22 (1), 55-69   Iqbal, M U., & Lim, S (2006) A Privacy Perserving GPS-based Pay-as-You-Drive Insurance Scheme International Global Navigation Satellite Systems Society Symposium, The University of New South Wales, School of Surveying and Spatial Information Systems Kelly, H (2012, October 30) Self-driving cars now legal in California Retrieved 2013, from CNN Tech: http://www.cnn.com/2012/09/25/tech/innovation/self-driving-carcalifornia/   Lanctot, R (2012, January 6) Insurers, Regulators Impeding Safety System Adoption Retrieved 2013, from Strategy Analytics : http://blogs.strategyanalytics.com/AMCS/post/2012/01/06/US-insurers-refuse-to-offerdiscounts-for-safety-systems-while-regulators-offer-only-praise.aspx   LeValley, D (2013) Autonomous Vehicle Liability-Application of Common Carrier Liability Seattle University Law Review , 36 (5)   Litman, T (2005) Pay-As-You-Drive Vehicle Insurance Implementation, Benefits, and Costs Victoria Transport Policy Institute Victoria, BC: Victoria Transport Policy Institute     -­‐  136  -­‐   Mikosch, T (2009) Non-Life Insurance Mathematics (Second Edition ed.) Copenhagen , Denmark: Springer-Verlag Berlin Heidelberg Murray, P (2012, August 15) Google's Self-Driving Car Passes 300,000 Miles Retrieved 2013, from Forbes: http://www.forbes.com/sites/singularity/2012/08/15/googles-self-driving-car-passes300000-miles/   National Highway Traffic Safety Administration (2014, March 31) NHTSA Annouces Final Rule Requiring Reear Visibility Technology Retrieved April 2014, from http://www.nhtsa.gov/About+NHTSA/Press+Releases/2014/NHTSA+Announces+Final+ Rule+Requiring+Rear+Visibility+Technology   Ohlsson, E., & Johansson, B (2010) Non-Life Insurance Pricing with Generalized Linear Models Stockholm, Sweden : Springer-Verlag Berlin Heidelberg   Pinto, C (2012) How Autonomous Vehicle Policy in California and Nevada Addresses Technological and Non-Technological Liabilities Intersect , (1)   Priddle, A., & Woodyard, C (2012, June 14) Google discloses costs of its driverless car tests Retrieved 2013, from USA Today : http://content.usatoday.com/communities/driveon/post/2012/06/google-discloses-costsof-its-driverless-car-tests/1     Witt, R C (1973) Pricing and Underwriting Risk in Automobile Insurance: A Probabilistic View The Journal of Risk and Insurance , 40 (4), 509-531         -­‐  137  -­‐