How discriminative is your car insurance?

The Consumer Federation of America (CFA) 2020 research has found that the car insurance premiums are 60% higher in predominantly black ZIP codes. Moreover, CFA has issued several reports showing that various socio-economic factors tend to raise insurance premiums for lower-income drivers, even those with clean driving records.

Based on extensive testing, CFA has determined that most insurance companies include non-driving factors to determine premiums, such as credit score, homeownership, marriage status, education, and profession.

Modern-day redlining in the USA

The term redlining roots from the 1960s and the discriminatory practice of Chicago banks literally outlining predominantly black areas in red ink to avoid investments (regular mortgage lending process). Nowadays, despite being illegal to ask drivers for a race, it's still perfectly legal for insurers to alter insurance premiums based on ZIP codes, justifying it by adjusting the theft and vandalism probability. As a result, the minority neighborhoods (immigrants, people of color) pay higher insurance premiums.

The majority of the US insurers still heavily rely on their insurance premium models on credit scoring. Over 50% of white US households have a FICO credit score above 700, compared to just 21% of black households. Similar numbers are observed in Hispanic households, resulting in disproportionately higher premiums for people of color. A growing number of US states are seeking to ban credit score-based car insurance premiums.

Homeowners, married, white-collar workers, and drivers with bachelor’s degrees or higher may expect lower premiums, shows the CFA report. At the same time, African-American home renters, singles, high school graduates, service workers, and drivers with average and low credit scores experience disproportionately higher penalties compared to white drivers. As per CFA, African-Americans are 4 times higher penalized for low credit scores.

Europe

Similar forms of non-driving-based discriminating factors are also observed in European countries.

German and UK insurance companies determine insurance premiums based on age, driving experience, driving record, car make and model, annual mileage, but also involve postal code and overnight parking location.

2018, British The Sun revealed that top insurance companies charge up to £1,000 more for car insurance premiums if the owner’s name is Mohammed, compared to John.

British GfK Mystery Shopping found that 21% of motorists above 80 years old were not even given phone or online quotes for car insurance.

In 2012 European Commission ruled on gender-neutral pricing in insurance.

Why does usage-based insurance matter?

Nearly every insurer's proprietary premium pricing model relies on various driving-related and a myriad of more or less non-driving and discriminating factors, such as age, sex, skin color, ethnicity, etc.

Besides the authorities combating the discrimination through a series of regulations, the telematics industry provides the solution.

Usage-based insurance (UBI) is considered the next best thing and the end of traditional car insurance. By tracking vehicle telematics data, such as speed, breaking, acceleration, the usage-based insurance better aligns with individual drivers’ risk levers. Telematics is what’s shifting the insurers’ discriminating models from socio-economic and non-driving factors towards precise claim risk factors.

Data about driving style, speed, driving time, etc., are collected using factory-installed or plug-in OBD-II port GPS tracking devices and smartphone applications. The critical part of risk identification is calculating the correlation between certain telematics data and vehicle crash statistics on a large scale.

Mireo SpaceTime Analytics

However, from the technological aspect, car data is extremely hard to handle at a scale. The time and geographical skewness of car data make it practically impossible to derive any data from it using conventional database solutions. As a technological enabler, Mireo, with its unique SpaceTime Analytics Platform, bypasses the gap between the telematics data and the insurers by offering a comprehensive analytical set of Big Data analytics tools.

Mireo’s SpaceTime Analytics Platform fuses the car data with the underlying digital map data, thus enhancing the car data to provide higher-order analytics, such as the hours spent driving on highways, within the cities, or above the posted speed limits. The unprecedented analytics speed enables actuaries to perform myriads of experiments to determine which country-specific parameters affect the crash probability and to what extent.

Mireo tested SpaceTime Analytics by examining trips from more than 500.000 Italian cars and their official crash report records to find how driving patterns affect risk probability. We ran a series of correlation tests and found both, some expected and some indeed fascinating results.

1. Italians driving between 23:00 and 04:00 have double the probability of having a crash.

2. Moreover, Italians spending more than 50 hours monthly on the roads are twice as likely to crash than regular commuters who spend 20 or fewer hours monthly on the road.

3. Frequent speeding, contrary to common perception, does not increase the auto insurance claim risk probability that much. You have to be a frantic driver to increase the risk of collision with speeding significantly.

If you've found these analyses inspiring and would like to see the similar numerous, on-the-fly calculated analytics running live on more than 200.000 vehicles, check out our free SpaceTime online demo.

Next steps

As insurers increasingly adopt telematics to offer accurately priced car insurance data protection and privacy are the burning questions. The Consumer Federation of America (CFA) considers usage-based insurance as a tool for eliminating the non-driving and discriminative factors in insurance premiums pricing. However, to avoid creating new forms of discrimination in auto insurance, the CFA advises state regulators to adopt certain consumer protections. Those include data usage consent, prohibition of data reselling, mandating insurers demonstrate the actuarial basis for the collected data, etc.