The Autonomous Vehicle Revolution
Thoughts on the timing and acceptance of automated vehicles
Introduction
It is no accident that the rapid rise in living standards over the past century has coincided with tremendous technological progress throughout the economy. At a fundamental level, rising living standards require the economy to generate an ever-increasing supply of goods and services through intelligent utilization of the fundamental factors of production: land, labor, and capital. Improvements in technology are reflected in the quality of capital goods, the ability of individuals to be more productive through enhancements in human capital, and better uses of limited natural resources. Progress on a per-capita basis is required for individuals, on average, to see growth in the economy translate into a higher standard of living.
Given the importance of technology in the economy, it is interesting to read Warren Buffett’s frequent statements regarding “not understanding” technology. He has expressed this sentiment in Berkshire Hathaway annual reports for many years as part of a discussion of the company’s acquisition criteria. However, what he appears to be saying is that Berkshire does not have a competitive edge when it comes to acquiring technology companies, not that Berkshire itself is somehow immune to technology or does not seek to adapt to technological progress. This reality seems to have gone unnoticed among some value investors who repeat Mr. Buffett’s assertion regarding technology and take it as a license to totally ignore the impact of technology on the economy in general and on their investments in particular.
The Opportunity
According to the Bureau of Transportation Statistics, there were over 260 million registered vehicles on the road in 2014. Currently, almost all of these vehicles have no autonomous features and are controlled by human drivers. The amount of time and energy spent controlling vehicles on the road is obviously very significant. According to the U.S. Census Bureau, it takes the average worker 26 minutes to commute to and from work which implies that the 139 million workers in the United States collectively spent 29.6 billion hours traveling to and from work, the vast majority in a private automobile controlled by a human driver.
Driving a two ton vehicle on public roads is something that does not lend itself to multitasking, at least not safely. Fully autonomous vehicles adopted throughout the economy would free up all of this time for other endeavors, including work, and would very likely cut the amount of time spent actually commuting because roads could be optimized more intelligently. In addition, the number of accidents on the road would theoretically plummet due to the decline of unpredictable human drivers. This could easily save thousands of lives annually along with significant monetary savings due to decreased frequency of property damage and bodily injury.
Elon Musk’s Tesla Motors is the most visible proponent of automation, although many other manufacturers and technology companies are also investing heavily in research and development. In 2016, the number of S.E.C. filings mentioning autonomous vehicles surged dramatically. Many of these filings are from automobile manufacturers but there is also increasing awareness of the benefits and risks of automation in other industries such as automobile insurance. Is this increased interest just hype or does it reflect reality?
A Question of Timing
If we fast forward fifty years to the mid 2060s, it seems almost inevitable that the United States economy will have fully transitioned to autonomous vehicles and reaped countless productivity benefits. This might seem like a bold assertion. Is this science fiction similar to the vision of the future from a 1960s Jetson’s cartoon? Even George Jetson’s flying car was not fully autonomous, although it did compress to the size of a briefcase.
Warren Buffett recently predicted that autonomous vehicles will eventually dominate the market but was skeptical regarding some of the more aggressive claims on timing. In particular, he was doubtful that even ten percent of the automobiles in the United States would be self-driving within ten years. Of course, the timing will be very important for GEICO as well as for other auto insurers. A much safer fleet of vehicles will reduce the rate of accidents resulting in less damage to property as well as bodily injury. This will cause insurance rates to fall significantly.
Even if we assume that a fully autonomous vehicle will be released imminently, it could take many years for the technology to become mainstream. The average age of light vehicles in the United States rose to 11.6 years in 2016. Furthermore, vehicles sixteen years and older are expected to grow from 62 million units on the road today to 81 million in 2021. Vehicle reliability has been increasing in recent years allowing consumers to keep older cars on the road without incurring major expenses.
The counterpoint is that perceived vehicle obsolescence will dramatically increase as soon as a fully autonomous vehicle is introduced. In fact, the comparison between the existing fleet and fully autonomous vehicles might be completely invalid. While the older vehicles will still get drivers from point A to point B, they will always require human control. The perceived difference could eventually be as great as the difference between a horse drawn carriage and an early automobile.
Flashback to the 1990s
Anyone who lived through the early days of the internet will recall the lofty predictions regarding how the technology would completely change our lives within a few years. These predictions were necessary to sustain the dot com bubble in which all sorts of dubious business models attracted capital, both from the unsophisticated and gullible and from those hoping to take advantage of the “greater fool” theory. Bricks-and-mortar retailing would soon be dead, we would all order products from our living rooms using computers, and everything under the sun would be handled online. These were just some of the more common predictions.
What ended up happening? Almost all of these predictions eventually came true! The only issue is that it took an additional ten to twenty years before technology and infrastructure reached the point where a critical mass arrived and existing business models could be seriously disrupted.
From an infrastructure perspective, we needed much faster internet connections than we had access to in the 1990s. Amazon.com perfected the logistics and infrastructure associated with massive automated warehousing and distribution. The smartphone revolution put miniature computers in the hands of nearly everyone along with software that increased price transparency and exposed retailers that were not offering good value propositions. Finally, consumers had to slowly get used to the concept of shifting more and more of their lives online. We are now at the point where an entire generation has grown up in a connected world and, of course, older generations eventually go along for the ride.
Change is Slow and then Fast
The impact of the dot com bubble on retail took a very long time to fully develop but it seems like we have recently arrived at a tipping point where the predictions of twenty years ago are about to come true. Malls have been in long-term decline but are now increasingly being razed and redeveloped. Sears Holdings has been declining for years and recently added “going concern” language to the company’s S.E.C. filings reviving concerns of near-term bankruptcy. Smaller retailers are often in even worse shape and quarterly earnings disappointments are common. It is still premature to suggest that bricks and mortar retail is dead but companies without pricing advantages or differentiation seem destined to fall victim to Amazon.com and others operating online.
Could we see a similar outcome for automated vehicle technology? The hype over vehicle automation is currently reminiscent of the online shopping chatter prevalent in the late 1990s yet we face significant technological and infrastructure challenges that will impede adoption in the near term. The technology required to safely automate a vehicle is still very much in development and unproven. The physical infrastructure of the United States is in poor shape, although the Trump administration claims to have plans to significantly boost infrastructure investment. Perhaps most importantly, people are used to controlling their vehicles directly and most people believe that they are “above average” drivers. Some people even enjoy driving.
In 2016, a high profile fatality in a Tesla vehicle resulted in a great deal of media coverage. Even though the ensuing investigation revealed no safety defects in Tesla’s autopilot technology, this incident and similar future incidents will slow down the shift toward automation both from a regulatory and consumer acceptance standpoint. As flawed as human drivers are, most of us feel like we are “in control” when driving. Handing over control to technology is a major shift that many people will resist. There will be a transition period when some vehicles are automated while others are human controlled. The interaction between autonomous and human drivers will also cause controversy, regulatory review, and impact consumer acceptance.
Conclusion
We have no way of knowing how the shift toward vehicle automation will evolve over time but the end result is easier to forecast. At some point, all vehicles are likely to be automated and will interact with each other in predictable ways. This will reduce the rate of accidents and boost human productivity as the time spent driving cars will be redirected to higher value pursuits. The benefits are likely to be immense but certain industries such as auto insurance will be much smaller as a result. As is the case with most technological change, there will be clear winners and losers.
What we cannot know at this point is whether the shift to automation will be largely complete by 2030 or if 2050 is a more realistic target. This has important implications for a business like GEICO or Progressive. The intrinsic value of a company is the analyst’s best estimate of the present value of all future cash flows from the business. If the decline of auto insurance is 30 to 40 years in the future, current valuations will be far less impacted than if the decline occurs in 10 to 20 years.
GEICO and Progressive have been growth companies for a long time as they took market share from competitors and benefited from the overall growth of cars on the road. At some unknown point, this virtuous cycle will reverse. Premium volume and float will decline. Companies will compete fiercely over the remaining business. Auto insurance companies will be valued as melting ice cubes rather than growth companies. Fifty years from now, we will know how everything turned out.
Disclosure: Individuals associated with The Rational Walk LLC own shares of Berkshire Hathaway, the parent company of GEICO.