By now, you’ve probably heard some of the hype surrounding self-driving cars; from Elon Musk’s vision of the future to the numerous concerns from lawmakers. The autonomous car will drive down the costs of transportation as we will no longer need people behind the wheel of taxis, Ubers, or buses. The technology will also eliminate the need for parking garages since drivers can send their cars home when not in use. While that seems like a distant vision; carmakers are surprisingly close to bringing autonomous cars to market, and for business owners and insurers alike, they couldn’t come sooner.
If your business has any sort of fleet of autos you are probably keenly aware of the insurance industry’s ongoing struggle with commercial auto insurance. Insurance companies have been consistently raising auto rates across their commercial books and businesses, especially in the transportation industry, have felt the effects.
There are several factors as to why the commercial auto insurance industry has been performing so poorly, but one stands out: the drivers. You can point to more miles being driven, rising costs of litigation, even how the more advanced technology used in cars increases the costs of repairs but all of those would be moot if we had a population of perfect drivers who never got in accidents. Like a ship’s Captain, the driver bears the ultimate responsibility of the car and is thus responsible for the auto insurance woes.
Theoretically, self-driving cars should be safer than those manned by human drivers for obvious reasons. Computers don’t get tired, lose focus, drink or get road rage; however, we all know that technology doesn’t always work in practice as it should in theory. We’ve already seen the first fatal accident involving a self-driving car. Plus, the autonomous cars will no doubt be sharing the road with other human drivers, adding a layer of unpredictability.
Insurance is fundamentally grounded in statistics. Insurers compile a tremendous amount of data on their policy holders, and use that data to infer the probability other policyholders will get in an accident, and how much it will cost. That is at the core of how rates are determined; past data dictates that a driver of a certain age, driving a specific make and model of a car in a given region has X probability of being in an accident. It’s complex, confusing, and imperfect, so insurers are constantly reviewing their data and tweaking algorithms in order to operate at a profit. This is still the case with nearly one hundred years and hundreds of millions of drivers worth of data.
By virtue of their business model, insurers can’t rely on anecdotal evidence that self-driving cars are safer; they’ll need lots of hard data. Without credible information on how self-driving cars act on the road, insurers will likely have one of two initial reactions.
The first option for insurers is to decline to offer coverage of autonomous cars, either by declining applicants with those types of vehicles, or adding wording to their auto policy specifically excluding coverage while the autonomous function is on. Currently, the auto policy has language specific to who is driving the car for coverage to be in force. It will be interesting to see how the courts interpret it in regards to autonomous driving. In case of an accident, is the person behind the wheel technically driving, or was the computer the legal driver? Insurance carriers may argue that if full control is given to the computer, their customer wasn’t the one driving. The policyholders, on the other hand, would argue the opposite in what could be an important precedent-setting lawsuit once it happens. Courts typically side with the insured in matters of policy interpretation but there’s no way to predict what the legal decision will be right now. In order to avoid costly legal battles, insurers will use specific policy wording to very clearly state their intention to not cover accidents while autonomous function is on.
Alternately, some insurers may respond to the risks associated with this new technology by raising rates. In the absence of the volume of data they need for ratemaking algorithms, insurers will initially err on the side of caution and estimate higher accident probabilities. Over time, if autonomous cars truly are safer, insurers should lower their rates commensurately. The originally skeptical insurers may then join the market after compiling credible data on the safe operation of autonomous cars, thus further lowering rates.
Either way, the most dangerous part of a car (the driver) is essentially removed from the equation and six-figure auto liability claims may become a thing of the past. My prediction is that the failure of autonomous technology will be treated as a manufacturer’s defect, much like if the brakes failed, and claims will shift from the driver’s auto policy to the car-makers Product Liability policy. This reduction in Auto claim payouts will result in rates stabilizing and Auto Policies once again becoming profitable for insurers and affordable for insureds.