Erie Insurance: Monitoring technology in the car insurance market and the issue of data privacy

Data and technology in the insurance industry

Comprehensive auto insurance policies provide broad coverage that addresses most common events a policyholder would face. They use the policy of “pooling and spreading” (Heide, 2020). However, when choosing insurance coverage, users must ensure they understand the fine print and what is covered or not: there are examples such as the woman whose tire exploded as she drove down the road doing damage to the wheel, rim, and body of the car, and her insurance ruled that since she had not had a collision or accident, the incident fell under an “act of God” and was not covered by her policy as written. Extra coverage may include protection against “fire, natural hazards, theft, windscreen repair, and legal expenses, among some other coverage that apply to specific events” (Gómez-Déniz and Calderín-Ojeda 2021). One scenario Erie Insurance specifically mentions is available coverage for pets travelling in the car. How companies offer these add-ons vary as they create bundles or stand-alone additions to a standard policy.

Insurance companies have always required large amounts of personal user data in order to best assign a level of risk and price point to their products. Historically, premiums have been calculated using individual characteristics, “such as gender, age, years of validity of the policy, etc.” (Gómez-Déniz and Calderín-Ojeda, 2021). Other factors may include vehicle’s age and model, marital status, and place of residence (Dijksterhuis et al., 2016). Increasing technology has increased the amount of data available on a given individual. There are now “millions of variables in multivariate analysis” (Carfora et al., 2019). Tracking client behavior leads to more accurate pricing and reduction in cost, as well as the personalization of premiums (Meyers and Hoyweghen, 2020).

With the introduction of monitoring technology, insurance companies now have better estimates of a driver’s risk factors. One tracking system, called In-Drive, provides information to the firm that goes beyond the driver’s address and total number of miles driven, along with a record of past accidents and claims or citations, which was formerly standard (Chen and Jiang, 2019). Many insurance companies have a large number of customers with zero claims. However, this does not necessarily reflect good driving ability. Some may not drive frequently or far, reducing risk in that manner, and some may not have claimed past accidents that were small in order to avoid an increase in their premium (Guillen et al., 2019). Now, by introducing telematics data, explain Chen and Jiang, the company has access to specifics, all recorded with GPS and mapping technologies, such as:

• The driver’s daily driving patterns including total number of miles and routes

Geyer et al. (2020) explain, “Each data point includes date, time, GPS coordinates, direction of driving, current speed, distance driven since the last data point, ignition status of the engine, and road type (urban, country road, or motorway).” This monitoring allows for “pay-as-you-drive” models of contract. Although one might assume that better drivers would choose this model for the possible savings, it is “more likely to be chosen by policyholders whose current Bonus-Malus rating is high (indicating high risk), and whose premium is more strongly affected by a potential downgrade” (Ibid.). There is a possible twofold benefit: Knowing they are being tracked can cause drivers to react differently and become safer (Guillen et al., 2019). Using data from car sensors leads to offers that are tailored to the individual driver, which would not be possible without the use of big data, beyond what can be gleaned from traditional methods (Lehrer et al., 2018).

Carfora et al. (2019) suggest that by applying machine learning and cluster analysis, there may be a way to compute driver aggressiveness based on location (urban or highway roads), since people often vary their driving style based on their environment. In the future, this may be used to place people into categories, such as “young driver,” “ruthless driver,” or “cautious driver” in order to assign a risk index. Weidner et al. (2016) suggest that the challenge lies in storing and processing these large amounts of information and extracting data from the variables that are relevant. Preliminary actuarial investigations have shown that telematics are effective but have not yet identified conclusively which variables are “precise and semantically correct” (Ibid.).

This is a newer segment of the car insurance market offerings, and a minority of the total written policies currently use this technology, but it is becoming more frequent in marketing and advertisements. First, we will discuss a particular insurance company and then the issues surrounding in-car monitoring for usage-based insurance.

Case study: Erie insurance

Erie Insurance, based in Erie, Pennsylvania, covers 12 states and the District of Columbia. Founded by H.O. Hirt and O.G. Crawford in 1925, Erie Insurance is now a FORTUNE^® 500 company with more than 6 million policies in effect for home, auto, and business. They have over 6000 employees at the home office in Erie and field offices and work with more than 13,000 independent insurance agents.

As of 2022, Erie is ranked the 13th largest auto insurer, based on premiums written, and is also within the top 20 for property/casualty insurance and business insurance. They were named to Forbes list of Best Insurance Companies in 2022. This may be due to their commitment to customer service, which has been recognized with twelve ACE (Achievement in Customer Excellence) awards since the awards were established in 2005. Internally, they have been recognized for flexible work options, encouraging healthy lifestyles and supporting diversity within their staff.

The issue of data and privacy

Common insurance practices require data in order to underwrite a client’s health and property. An increasing amount of data have led to an increase in sophistication of available risk models. With every advancement, some negatives always exist to balance the advantages, and in this case, it is the balance of providing information for savings versus maintaining privacy. Loi et al., 2022 explain how big data analytics and artificial intelligence contribute to the telematics employed by car insurance companies and that “these technical developments…allow dealing with risks that were previously considered uninsurable. …The availability of large amounts of data that can be used to assess, select, price, predict, and prevent risks is key.” As more and more customer data are made available for the purpose of risk modeling, some consumers may fear the loss of their privacy. However, Loi et al. claim that “insurance companies have an ethically justified need to obtain client data to calculate risk-adequate premiums and prevent moral hazard,” making it hard to define the limit.

Analysis of the traditional data provides valuable information. Factors that contribute to crashes, injuries and fatalities are divided into three categories: “those associated with vehicles, those associated with drivers (and pedestrians), and those associated with roadways” (Fowles & Loeb, 2021). Fowles and Loeb spell out variables that have been emphasized in the past, such as “motor vehicle speed, speed variance, alcohol consumption, speed limits, vehicle miles travelled, measures of income and wealth, the age of the motor vehicle fleet, seat belt usage and seat belt legislation, and the deregulatory climate of the 1980s,” and point to the impact of cell phones in more recent times as their use has increased exponentially, citing the numbers climbed from 340 thousand in 1985 to over 310 million in 2010.”

Statistically, it seems that policy holders are “more likely to encounter accidents during the last month of the insurance policy term than during any other month.” This “sunk cost fallacy”—the idea that policy holders feel the premiums have been wasted if the coverage is not used—causes them to reduce their “accident-prevention efforts” and increase their “moral hazard” (Ma, 2021). Although this does not seem to make sense from a practical standpoint, the numbers point to the phenomenon. Insurance companies continually search for ways to protect their customers and reduce the need for accident claims. One way to do this is with the mentioned monitoring technology, which gives the insurance companies a better view of the driver’s true activities. When implementing this technology, the volume of data challenges traditional statistical techniques. Guillen et al. (2021) recommend weekly telematics summaries are necessary in order to process the amount of information.

Cyber security

With the amount of data flowing through the insurance industry, companies must be extra vigilant about cyber threats and data breaches. At the same time, it opens the door for them to offer new products and services in this arena. InsurTechs, digital startups that allows companies to “improve and enhance their digital capabilities” is changing the way insurance companies work and interact with customers, generating more value for that customer (Pisoni, 2021). Insurance companies have “vast amounts of privacy-sensitive data” on their customers, especially when GPS information from tracking information is added to the equation (Derikx et al., 2016). Beyond the location data, these connected monitoring devices retain IP address, WiFi and router information. Most users seem to feel that the fair pricing and financial compensation of usage-based care insurance overcomes any hesitancy some may feel about the using an installed device or mobile-phone-based tracking app. A study by Derikx et al. (2016) showed that “consumers may actually attribute positive utility to relevant personalized advertisements from insurance companies based on car usage data” and appreciate more personalized and relevant offerings.

Telematics data are subject to data protection laws, as it records activities of individuals that may be identified. Naylor (2017) points out that ADA systems must also be factored in and sensors checked regularly to ensure that the data being recorded is accurate.

Benefits of monitoring technology

“The key elements of the insurance models remain the same,” says Pisoni (2021), but the industry must evolve to reflect the times. The increase in monitoring technology naturally leads to the idea of autonomous or self-driving cars. It is interesting to speculate what the future may hold, and this may also affect insurance companies and the way they provide services.

Naylor (2017) reported that car crashes kill more than 30,000 per year in the United States, and at the time of his article, KPMG reported that the implementation of networked autonomous cars would drop this rate by at least 80%. Even if auto-driven cars are not crash-free, the rate would drop because an estimated 93% of them are caused by human error and influence by physical, emotional, psychological, or intellectual factors like being tired, angry, confused, or distracted (Ibid). Tesla currently has semi-autonomous cars, and Volvo and other companies have concepts in development, with Samsung working on compatible technology. Because cars will require voice and face recognition to start and can be programmed to lock the doors and deliver a thief to the nearest police station, the risk of theft will also fall. Naylor concludes that the insurance company would then be more concerned with minor damage caused by weather or incidental factors like a shopping cart in a parking lot. Companies are still testing this technology and the industry will evolve around the platforms and programs that work best.

Conclusion

In order to succeed and continue to be attractive to consumers, monitoring technology must come with a level of trust. A company like Erie Insurance has a focus on customer service, loyal users who have been with the company for years, and a dedication to provide quality products and protect their customers. For these reasons, if they continue in their mission, they should do well adapting to the challenges of an increasingly technological market and offering attracting and competitive car insurance offers.

Questions

1. When it comes to car insurance, many people—young and experience drivers alike—do not fully understand all the options and walk a line between full coverage and affordability. What are your experiences with car insurance and what factors are valuable to you when choosing a carrier?