It is in the insurance industry’s best interest to use automated driving technology to make things safer through crash avoidance, not just crash mitigation. Unlike crash mitigation, crash avoidance saves the insurance company money. This was Alain Kornhauser’s premise for the Smart Driving Car Summit, Incentivizing Through Insurance. The trillion-dollar question is whether the incentives to automate will overcome the countervailing forces of existing business practices in other parts of the mobility industry?

Access to Accurate Data Is Key #

Bringing the European perspective to the discussion, Jacques Amselem, Head of IoT, Allianz Technology, a €140B financial services and insurance firm, believes the EU has “a proper regulatory framework to insure automated driving.” The key is that they have insurance instruments to protect the driver/auto and product liability to protect the manufacturer of the vehicle. The major issues he sees are:

1. Access to the data the car produces for understanding and preventing collisions to help improve vehicles as well as helping customers gain trust in the technology.
2. How does one provide insurance for something dynamic, like artificial intelligence, which is continually evolving (hopefully improving)?

AI is the most interesting part of the future of insurance, according to Michael Scrudato, SVP Strategic Innovation Leader, Munich Reinsurance America, Inc., an insurer of the insurers. Echoing Amselem, he indicates that the offering shifts from insuring drivers to insuring against product/technology/software defects.  Access to the driving and collision data is central to unlocking the cost reduction benefits.

For instance, timely and objective collision data removes friction, reduces time to resolution, and minimizes she said/he said questions in determining fault. It’s not a clear path, as some of the challenges include:

• How can the insurance industry get a better perspective of what safety technology is on a given vehicle? Simply, this information is not readily available.
• Scrudato also has concerns around the development, testing, certification, integration of automation in vehicles.

Still, Scrudato believes that “Insurance should not be a barrier to the adoption of automated driving technologies.” A 2020 RAND Corporation report (PDF) echoes Scrudato’s comments.

Kornhauser comes at this from the perspective of an entrepreneur and sees an enormous opportunity for a disruptor to get the buying public to purchase products that provide enormous social benefits. The real value is if the technology can prevent a crash.

Insurers Are Like Goldilocks #

There are 50 different state departments of insurance and they regulate rates like Goldilocks; not too high (to the detriment of consumers), not too low (path to insolvency). Because of this tightrope act, insurance companies have to have good data justifying things like rebates or discounts for safer technology. As a result, insurance companies have limited ability to use pricing signals to incent owners or OEMs to deploy ADAS. Diana Furchtgott-Roth, Adjunct Professor of Economics, George Washington University, observes,

“We need to change the behavior of the insurance company regulators at the state level before we can change the behavior of the insurance companies,”

States are starting to change laws to allow insurance companies to provide rebates for automotive safety technologies according to Scrudato and he calls this a good U.S. regulatory development.

Accurate data is key and Amselem believes it starts with higher transparency from OEMs. Unlike the U.S., where OEMs self-certify their safety technology, European regulators require third-party certification. Amselem prefers the European approach since it removes the potential conflict of interest associated with self-certification (the story of the fox guarding the henhouse) and it helps build trust.

He believes the certification entity should be publically-funded and not have to rely on private funding. The panel moderator, Compass Transportation’s Richard Mudge, agrees that some sort of public funding is necessary for the U.S. as well, “given the pace and scale of new cars.” 

To date, the U.S.’s answer is the insurance industry-funded, non-profit, IIHS. The IIHS provides an analysis of its testing of crash avoidance technologies and crash results for the public. It also analyzes past crash data to provide consumers with a picture of the relative safety and the insurance costs of various vehicle models. Still, currently, neither IIHS nor NHSTA provides independent certification.

Scrudato focuses on applying technology to improve insurance products, such as their Smart Mobility initiative. This program combines collision avoidance, telematics, driver coaching, and fleet monitoring to help fleet managers reduce crashes – thereby saving money and heartaches. He indicates that one of their tools, Lost Detect, which analyzes crashes, suggests that potentially (assuming the driver reacts appropriately) 2/3 of dollars lost to crashes could be prevented.

A Shift in Mindset #

This idea that crashes are preventable represents a shift in mindset and is still evolving, according to Scrudato. Scrudato cites the progress since 2014 as follows:

1. Availability of database insurance products that allow pay-by-mile insurance. There used to be a question of whether people would share driving data. Now, all the national carriers provide those products. Scrudato notes that new policies and new drivers are much more likely to take these mileage policies. It is also a gamechanger for parents, as they now can view their teenagers driving performance.
2. In the commercial fleet space, fleet managers are implementing data collection for safety and not just regulatory compliance. Scrudato says they are using data, whether telematics or camera, to find bad drivers. The data are the basis for custom driver training based on what individual driving habits are. Scrudato says that “Videos being pushed to drivers that they have to watch” if they want to continue being a driver. It seems to be creating a culture of safety.
3. ADAS is more widely available, particularly in the commercial fleet space than in 2014. This is a gradual and evolutionary change, but it is happening.

To some extent, this means technology simply becomes another tool for insurers and their clients to reduce the total economic and non-economic costs of crashes. Dr. Kara Kockelman of the University of Texas estimates the

“comprehensive crash costs are roughly $1 trillion (yes, trillion) per year in the US, so about $3k/capita/year (far beyond costs of congestion & such).”

This equates to approximately $3,000 per capita per year or approximately $0.10 per mile (simple average), according to Kockelman, who is co-author of a 2018 paper (PDF) that examines these costs and the potential for automated and connected safety technologies to ameliorate their economic and human impact.

Perhaps The Disruption Starts with a Safe Bus #

Insurance pools are an alternative for fleets. One such entity is the Washington State Transit Insurance Pool (WSTIP), which, as its name implies, is a means for public transit agencies to share risk and improve safety. Jerome Lutin is leading a WSTIP research project that is examining the potential cost savings associated with adding technology to make buses even safer than they are today.

And although buses are safer than cars, the costs are significant as Lutin cites Federal Transit Administration and National Transit Database data suggesting that between 2014 to 2019 the cumulative Casualty and Liability expense was almost $4B. These costs don’t count the personal trauma associated with the 32k+ collisions, 99k injuries, and 583 fatalities over that time.

The demonstration project has equipped 30 buses with Pedestrian Avoidance Safety Systems (PASS) and Lidar-based forward collision and pedestrian collision avoidance warning and braking. Some of the subtle things they are examining, beyond the effectiveness of preventing crashes and associated damage, is the impact of automatic emergency braking on bus passengers (who don’t have the option of a seat belt or are often standing).

They are in the data-collection stage of this five-phase project. Their data promises to be a valuable tool for the industry, as it will be made available to other researchers.

Lutin points out that for transit systems, the average annual cost of insurance claims is about $6k per bus. He estimates that approximately 75%, or $4,500 per year, is the collision insurance portion of this cost. With a typical 12-year bus life, this provides an upper boundary in terms of how much can be invested in technology in order to reduce insurance costs.

How to Overcome Moral Hazard? #

What the WSTIP program already demonstrates is that a fleet operator has a major built-in incentive to reduce risk by improving safety. What they are selling is mobility and not, as Dr. Kornhauser would say, “the fins and chrome.”

An obvious issue in the consumer market is that automated safety features represent significant upfront costs, which act as a barrier to adoption. But somewhat hidden is the high cost of repair of these new features. For instance, Amselem points to a 2-year study of cars equipped with ADAS and found out that the replacement cost of a windscreen (windshield) with integrated stereoscopic cameras is ten times higher than a traditional glass windscreen.

So, although the number of crashes may decrease, the total costs to repair a crashed vehicle increase. This is reflected in an increase in insurance rates for cars with these advanced features. As a real-life example, Mudge points out that his insurance rates increased when he purchased a Tesla, even though it is presumably safer than his previous vehicle.

OEMs and their dealer ecosystem have little incentive to reduce the cost of repair. As Kornhauser says, “Crashes drive new car sales and repairs.” Crash avoidance doesn’t drive new revenue. Mudge correctly points out that consumers will have to demand vehicles with relatively low-cost life-cycle cost safety features. The question is whether OEMs will be able to respond to such a demand without sacrificing other customer wants around things like style, comfort, or efficiency?

This isn’t to say that things aren’t getting better. From his vantage point as one of the early thinkers about using automation to improve safety, Scrudato says that we have seen, “substantial improvements.” He “doesn’t think the way insurance works has to dramatically change for Level 4 vehicles.”

And given that the early rollout of Level 4 is going to be in the form of driverless vehicles, perhaps the disruption will come from this emerging transportation sector. Since the incentive of MaaS (Mobility as a Service) is to provide mobility at the lowest cost deploying vehicles that minimize total lifecycle cost will be a priority. And, using Lutin’s estimates there are potentially significant insurance savings by implementing automation in a fleet setting. Simply, the driverless fleet of the future is about selling a service, as opposed to selling products like traditional OEMs.

Stay tuned, as next week, 3/11/21, the Smart Driving Car Summit will delve further into the role of regulation both for self/safe driving and driverless.

Note: The author acknowledges and appreciates the feedback from D. Kara Kockelman and Diana Furchtgott-Roth in helping improve the clarity of this summary of the 3/4/2021, Smart Driving Car Summit session.