The technology that will enable us to get around in self driving vehicles is closer to reality than we think. However, mass adoption of fully autonomous vehicles will be further away than the markets anticipate. Car insurance could become an obsolete product somewhere between 2044 and 2060, predict the authors of an in-depth study.

Entitled Insuring Autonomy: Analyzing the Implications of Self-Driving Cars for the Auto Insurance Industry, the study was published early last autumn by Morningstar DBRS. Its authors are Brett Horn and Suryansh Sharma.

With regard to road accidents in the USA, per mile driven, the number of accidents and fatalities declined by almost 70% between 1980 and 2010. However, this progress has stalled, and there has been a modest increase in recent years.

Human error is the main cause of almost 90% of crashes. “Human driving habits and behavior are notoriously difficult to change,” say the authors.

Caution  

The report suggests that investors take a cautious approach to insurers’ stocks. Among the insurers most exposed to auto risk are Allstate, Berkshire Hathaway, Progressive and Travelers.

“With the group trading at a hefty premium to historical book multiples, from a long-term perspective, we question whether current valuations are justified for businesses that might become obsolete.” 

Over a period of 10 to 20 years, the value of auto insurance portfolios will have to be written down, as policies will no longer carry a deductible payable by the driver.

Markets have been excited by the prospects of this emerging market for years. The median capital-raising value of startups developing technologies for autonomous vehicles (AVs) has increased sixfold between 2019 and 2024.

Five tiers  

There are six levels of autonomous driving, from level 0, where assistance is momentary, to level 5, where the vehicle is fully autonomous and requires no human assistance. Levels 0 to 2 still require driver supervision or intervention.

Almost all the most advanced technologies fall into levels 2 to 4, with the majority in the second tier. Only the Waymo robotaxi service appears in level 4.

Waymo has published an analysis with Swiss Re, reported by Morningstar DBRS, in which no accidents with injuries are reported for robotaxis that have covered 3.8 million miles (6.1 million kilometers). In comparison, drivers reported an average of 1.1 claims per million miles travelled for accidents with injuries.

For vehicle damage, the claim rate reaches 0.78 claims per million miles, compared with 3.26 claims for vehicles driven by people.

“The entire point of autonomous cars is that the driver is no longer in control. Therefore, it is hard to see how the owner of a fully autonomous car could be found legally liable for damage. In our view, car insurance would most likely morph into product liability insurance and would ultimately be borne by the auto manufacturers or the company that provided autonomous capability in the vehicle,” say the authors, adding this would likely happen when level 4 or 5 autonomy is reached. 

Until then, as long as human intervention is required, personal insurance for the driver will still be necessary. “We think multiple hurdles will need to be overcome before drivers are no longer required to maintain insurance. First, highly autonomous vehicles will need to be available on a wide scale. Second, regulations around insurance will need to be changed on a state-by-state basis. Finally, autonomous vehicles (level 4 or 5) will need to meaningfully replace nonautomated and even partially automated cars on the road.” 

Three scenarios 

The study analyzes the gradual implementation of driver assistance technologies in three scenarios: very aggressive, aggressive, and moderate scenarios. The time taken for the automotive industry to adopt these technologies was compared, as an indication of the time required to reach autonomy levels 4 or 5 in 80% of vehicles. Some 10% of vehicles are already equipped with driver assistance systems.

To go from 10% to 80%, depending on the scenario, will take between 7 and 18 years, the authors estimate. To reach 60% of AV on the road with level 4 or 5 automated systems, the very aggressive scenario will achieve the target in 2044, compared with 2053 for the aggressive scenario and 2060 for the moderate scenario.

The authors add that we must not overlook the impact of the resale of vehicles that are not equipped with driver assistance systems, which will be abandoned by their owners in favor of AVs. Instead of being repurchased by a new driver, a higher proportion of these vehicles could be sent directly to the scrap heap. 

The impact on insurers will occur in three stages. According to the very aggressive scenario, for an initial period of 10 years, there will be no impact on automobile insurance products. Even if the adoption of AVs and their level of autonomy progresses, normal delays in the required adjustments to laws and regulations will mean that insurance will be maintained, even if it is no longer needed.

Thereafter, for another 10 years, as AVs are adopted by consumers, insurance premiums will fall at the same rate, forecast the authors. They also assume “underwriting results weaken due to cost deleveraging, pushing returns toward the cost of equity.” 

“There is potentially a positive scenario for auto insurers. If full autonomy proves difficult to achieve but partially autonomous cars meaningfully improve safety, auto insurance could remain necessary for decades to come,” says the report. 

“Auto insurers could potentially draft off lower accident rates for years. Historical results suggest auto insurers produced stronger returns when accident rates were falling before 2010.”

Competition 

The study also assesses which countries are most advanced in terms of autonomous driving. Although Europe has some very advanced technology companies in certain niches, the Morningstar DBRS researchers compared the chances of China or the USA in this race for dominance.

The researchers use six criteria to compare the countries, in declining order of importance: the software stack (which replaces the driver’s brain), competitiveness and dynamism, governmental and policy support, the hardware stack (the devices that replace the driver’s eyes), ecosystem maturity, and consumer acceptance.

On the software development side, the USA is clearly ahead, but China dominates in several other areas, notably competitiveness and social acceptance.

In other areas like policy support, regulatory frameworks, and the overall maturity of the autonomous vehicle ecosystem, it’s not yet clear which country is ahead of the game, according to the authors. They do, however, give China a slight advantage.

Numerous and varied impacts 

This article summarizes the main highlights of the study, but the very detailed document runs to 81 pages. It presents the current status of various driver assistance systems. It also reflects on the usefulness of insurance in a context where the roads are traveled by a large majority of AVs.

“A way to think about job disruptions from autonomous vehicles is to think about the direct impact on jobs that mostly involve driving motor vehicles, such as truck drivers, bus drivers, taxi drivers, and so on, as well as the indirect impact on jobs that somewhat involve driving, such as firefighters, electricians, construction workers, and so on.” 

Once the risk of collision has been eliminated, all that remains is to protect the vehicle against damage caused by something else, including weather events such as flooding or hail, or theft. Taking Allstate as an example, according to which catastrophes account for 2% of claims over a five-year period, the authors point out that “if auto insurance coverage were limited to this area alone, the industry would shrink to a negligible size.” 

Liability related to driving the vehicle will become more complex to determine when the autonomization process reaches levels 3 and 4. At level 3, the driver can still intervene. At level 4, liability will fall to the vehicle manufacturer or its suppliers in the event of a collision, reducing the contribution of traditional insurers to very niche products.

Update  

A new update of the joint Waymo/Swiss Re study has also recently been published, based on the behavior of robo cabs that have travelled more than 25.3 million miles (40.7 million kilometres) fully autonomously. Entitled Do Autonomous Vehicles Outperform Latest-Generation Human-Driven Vehicles? A Comparison to Waymo’s Auto Liability Insurance Claims at 25 million miles, the study was published in December 2024.

After comparing the data obtained via Waymo to all American drivers, the researchers report an 88% reduction in property damage and 92% reduction in bodily injury liability claims. Waymo reports 10 collisions, 9 of which resulted in property damage and 2 in bodily injury.

Even when comparing the robotaxi’s results with newer, technologically advanced vehicles, the reduction in claims is almost the same.