Statistics undergrad students investigate insurance for self-driving cars

June 12, 2019 by Jessica Macy Lam

How will our roads look like in ten years? According to undergraduate students, Kangdi Yang, Yaru Yang, Yuxuan Zhang, and Qiwen Zhao, in the University of Toronto Department of Statistical Sciences, self-driving vehicles are likely to take over. And once autonomous cars hit the roads, how are we going to insure them?

For a student research competition hosted by the Society of Actuaries, the team of students, otherwise known as Team InnoVision, wrote a research report on how autonomous cars will likely change the insurance industry – and insurance policies for car owners. Each student team received data on a fictional city’s number of insurance claims for traditional vehicles and claim amounts.

With the help of Professor Sam Broverman, who had introduced the competition to his students, the team embarked on a month of intense data cleaning, analysing and programming. Team members also brushed up on market research, as well as the various factors affecting insurance premiums, which helped them get ready to forecast the terms of tomorrow’s car insurance, as well as future car insurance for traditional vehicles, in the case where both types of vehicles are on the road together. (The team’s submission made it all the way to the competition’s semi-final.)

One of the students’ biggest findings happens to be very good news for owners of self-driving cars: car owners likely won’t ever have to worry about being blamed for accidents again, as self-driving cars take human error on behalf of the driver out of the equation.

Crashes involving vehicles we use today are largely caused by people. A crash involving self-driving cars, however, would be caused by system failure, technology failure, or network failure, the team says.
 
“[Car crashes] is something the manufacturing company would be responsible for but not the driver, that is the kind of transformation we need to take into account and abandon some of the old risk factors,” Qiwen Zhao says.
 
So, what is going to change in the insurance industry once self-driving cars officially hit the market?
 
A possibility is that consumers may not need to pay if they are in a collision in a fully autonomous vehicle. Everything may be covered; however injuries to a car’s occupants could possibly be more severe depending on the size of the vehicle and the fact that high-end cars are usually equipped with sophisticated technology. Since the car would rely on technology, any failures could result in possible collisions that could injure the occupants.
 
Actuarial science, which assesses risks in insurance and finance, will play a huge role in the future of self-driving cars, Zhao says.
 
“You always have insurance for whatever vehicle you’re driving,” Zhao says. She adds that any collisions that happen because of errors made by manufacturing companies, the manufacturer would take the losses instead of the insurance company.

Actuarial methods help predict what the future insurance premium should be. In this case, the team found that the price of the insurance premium would decrease because the more fully autonomous vehicles on the road, the less accidents are likely to occur.
 
Yaru Yang points out that changes to today’s car insurance industry and policies could be a reality in the not-so-distant future. The arrival of fully self-driving cars to the market might happen a lot sooner than most people think and some autonomous vehicles with technological assistance are already here.

“[Insurance industries] need to consider now that this technology is already on the market. Companies are going to compete for these market shares and the companies who have a solution in place, will win,” Yang says.

Without proper research and preparation, companies could make the mistake of making premiums too expensive, leading consumers to choose other companies who have lowered the premium. Even if fully autonomous vehicles are not yet on the market, Yang says, insurance policies need to be designed today.

“Thinking about it early is the most important thing.”