Google has announced a truly “driverless” car that lacks a steering wheel, brakes or anything else that allows a driver to control its movements. This actually strikes me as a much bigger deal for the insurance industry than the optionally self-driving cars that have received so much attention and seem likely to come to market within the next few years. And, while the short-term insurance implications of the type of self-driving cars that are coming to market soon seem pretty modest, the new truly driverless models seem a lot more likely to have big effects.
Let’s start with what’s going to happen soon. Self-driving cars that require a “pilot” in a driver’s seat and have a steering wheel, accelerator and brakes are really just an evolutionary change in vehicles that progressively have become more automated over the last decade. As such, the changes in auto insurance are also going to be evolutionary. All major automakers already sell models with traction control, self-parking, collision avoidance and adaptive cruise control features. Together, these features make a car “half self-driving” already.
The “self-driving car” that will probably be on the market within a few years just adds a steering and navigation functions to this already widely deployed suite of features. Since drivers will still be able to control these cars, they will still need liability insurance. Some claims will become product liability claims—some types of accident claims already are—but most accidents would still result from human error of various kinds. There isn’t going to be a huge change. The insurance business will continue in much the same way for most insurers and most consumers.
There’s one exception: self-driving cars, even in the early generation, will be nigh-impossible for street criminals to steal. Any self-driving car is almost certainly going to be traceable via its GPS system and have a “kill switch.” Protection against theft isn’t a huge part of auto insurance premiums in most places but, in the long term, self-driving cars seem likely to more-or-less eliminate an entire category of crime. (Computer hackers might still steal cars from time-to-time but I’d suspect this is going to be about as difficult as raiding a bank account, which is hard.)
However, the kinds of fully self-driving cars that Google is now testing could represent a much bigger change. If a driver can’t manipulate a car in any way (except maybe to press a “stop” button) most crashes will probably result in product liability claims. Although such product-liability auto insurance could, in principle, be purchased as “master policies” by automakers, the mere fact that we have a longstanding cultural habit of buying auto insurance makes it quite possible that consumers will still buy polices. And, most likely, automakers as well as some insurance and consumer groups will argue for offering these policies on a no-fault basis. This, in turn, could lead to a real resurgence in no-fault coverage that’s fading elsewhere.
Second, under a full-self-driving model a significant fraction of people may well move towards a fractional ownership or “car-sharing subscription” service. Car sharing, of course, is already reasonably widespread in dense urban areas but remains a niche market because it’s not economically efficient to use car-sharing to commute to work, go on a road-trip or, really, do much of anything besides run a brief errand. By contrast, a service that lets you summon cars without drivers to wherever you are and have them take you where you want to go could replace automobiles for many people. One would think these services would probably bundle in some sort of policy in just the same way that existing car-sharing services do.
The near future of self-driving cars probably isn’t a big deal for insurers. The more distant future of fully autonomous cars, however, may well result in big, big changes.