Driverless Car (Part 4): Will Insurers Survive?

Before concluding this series on driverless cars, I’m going to take a detour and use this column to delve into the cars’ potential impact on the auto insurance industry. While I’ve already mentioned the issues in passing, they are starker and more imminent than most realize and deserve a deeper look. The doomsday scenario is clear for the roughly $200 billion in personal and commercial auto insurance premiums written each year in the U.S. Insurance premiums are a direct function of the frequency and severity of accidents. In a world of driverless cars, where accidents are significantly curtailed, most of those premiums will go away. Sure, some car insurance will be needed, but the market might be reduced by 75% or more. Insurers make their profits on the float from their premium income, so plunging premiums spells doom for many insurers. However, based on numerous conversations, it is clear that insurance-industry executives mostly just roll their eyes if asked to contemplate the implications of driverless cars. Even if the driverless cars are possible, conventional wisdom goes, it will be decades before they are relevant. Therefore, there is little need to worry now. Here’s how insurers figure the math: Begin with the assumption that it will be years before the technology matures. Add several more years to sort out the regulatory complexities, including licensing and liability issues. Add some more years to gain consumer confidence. Then, given the long lifespan of cars, add another decade or more before driverless cars make up a significant percentage of the cars on the road. On top of that, the argument goes, even if the frequency of accidents goes down, the severity will go up—as measured in the cost to fix cars with all the cameras, sensors, radars, etc., that are going into them. And, remember, even if you don’t crash into someone else, someone else might well crash into you. So, it will be decades before anyone could even imagine giving up car insurance. Besides, there might be no short-term cost to being wrong. Fewer accidents would just mean fewer claims, and therefore greater profits, until enough actuarial data proved that driverless technology delivered the conjectured savings and forced premiums down. Thus, the prevailing attitude is probably much like that of Glenn Renwick, CEO of Progressive Insurance, as expressed during Progressive’s February 2013 earnings call: The technology to do an autonomous car has been around for a while. We’re now seeing them; we’ll see a lot of talk about them. The real issue is exactly how they are able to be part of the fleet of vehicles on the road in America, and that is probably not something that need keep anyone awake for quite some time. In fact, it’s time for insurance executives to lose a little sleep over driverless cars. For one thing, far-off doomsdays have a way of sneaking up on you. As Paul Carroll and I documented in “Billion-Dollar Lessons,” Kodak concluded through very sophisticated market research in the 1980s that it would not be threatened by digital photography for a decade or more. It was right. Unfortunately, it did little to prepare for the inevitable disruptions. When it did attempt to mobilize, the advantages that it once held had little relevance. A succession of CEOs could not stem Kodak’s decline into bankruptcy. The more tangible danger is that Google’s driverless car program has started a technology arms race across the auto industry. If auto industry executives and boards of directors were not focused on this transition before, they are paying attention now. Most automakers are racing to differentiate their premium models with intelligent driver-assist functions like smart cruise control, accident avoidance and crash monitoring and reporting. These efforts will hasten consumer trust in driverless technology and accelerate the proliferation of the technology throughout all car models. As an example, Volvo, an automaker known for safety but relatively small in terms of global sales, predicts that it will be able to eliminate crashes altogether for anyone driving one of its cars by 2020. If tiny Volvo can aspire to this audacious goal, what might Big Auto be able to do? While, as I discussed in Part 3 of this series, this incremental approach might not save automakers from their eventual business model doomsday brought on by totally autonomous cars—it will hasten the disruption for auto insurers. Guy Fraker, a former director of enterprise innovation at State Farm Insurance and now an executive at AutonomouStuff, says most accidents happen in congested traffic. The accidents, in turn, cause more congestion and more accidents. Fracker argues that even a 25% adoption of incremental driverless technology such as smart cruise control and crash avoidance would significantly relieve congestion and reduce the number of congestion-related accidents. Fraker’s analysis is consistent with the forecast that one large insurer shared with me. That forecast estimated that a 20% adoption rate of incremental driver-assist technology might result in significant enough reductions in accidents to trigger material reductions in premiums. In other words, insurers will feel the effects of driverless technology long before fully autonomous cars become ubiquitous. Because premiums lag actuarial data, insurers will face strategic choices. Some insurers will delay instituting price reductions and enjoy greater short-term profitability. Others, more focused on the transition, will use the drop in claims to be aggressive on pricing, stealing the best customers and gaining market share. The industry may well have to start making these strategic choices in the next few years. Fred Cripe, a former head of product operations at Allstate Insurance and now an industry adviser, is more sanguine about the prospects for insurers in the short term. Cripe believes that adoption will be faster than many industry executives assume but that it will be more chaotic than predicted by advocates of driverless technology. He reasons that human drivers will become more erratic in the short term as they adjust to the technologies, sending accident rates up. Cripe agrees, though, that in the long term, “car insurance goes away.” He also says that, to prepare, the insurance industry needs to fight its tendency to push away liability. Insurers issuing policies based on driverless technologies will be tempted to set high prices or to simply refuse to cover driverless technology at all. The danger with this approach is that it leaves an opening for someone else to innovate and create the right insurance products and business models for the emerging world where driverless cars are pervasive. Cripe argues that, “rather than trying to push the losses out of the system,” insurers should focus on product innovation to “maintain control of the losses.” Google could use its deep pockets to offer cheap insurance, to hasten adoption and grab market share. Automakers could embrace the inevitable increase in their product liability and bundle insurance with their vehicles—rather than letting downstream players take those profits. Fraker hypothesized that automakers could grab a large chunk of the future insurance market by simply extending their warranty operations. Big Venture start-ups might accelerate the emergence of car-sharing fleets, thereby significantly lowering the need for personal insurance coverage. Or, in classic innovator’s dilemma fashion, start-ups like mileage-based insurer MetroMile might eat away at the edges of the market and be best prepared to take advantage as the disruption grows. Whatever the competition does, a traditional industry strength—underwriting—will decline in importance. For decades, insurers have invested heavily in their abilities to assess individual risk and price accordingly. As one CEO told me, “Insurers with better information make smarter underwriting decisions and slaughter those with less; that’s the nature of the business.” But driverless technology will take human error out of the equation and make underwriting less important. The basis of competition will shift to other aspects of the business, such as customer relationships, claims processing, expense management and distribution. The shift will create openings for new competitors such as CoverHound, a well-funded San Francisco startup that has built a multiple-carrier comparison engine. Basil Enan, Coverhound’s CEO, believes that auto insurance will become much more like homeowner’s insurance, where claims are rare but very high. His view is that auto premiums will, of course, go down but that claims and other operating costs will, as well, so insurance will be a more profitable business for the survivors. He plans to manage the insurance relationship for driverless technology adopters by helping them sort out the right policies.