We think autonomous vehicles will be one of the most dramatic changes to humanity this century. The move to AVs will have many ramifications in and out of the automotive industry, as not only will vehicle interiors and design change, but so will the manner in which people use their new free time because they eventually will not have to watch the road. Over the next decade, we see the industry moving toward Level 4 autonomy, in which vehicles operate without a steering wheel or pedals within a geofenced area, most likely in dense city centers. We expect these vehicles will operate as part of fleets managed by ride-hailing firms such as Uber and Lyft, as well as by ride-hailing operations yet to be formed by automakers. We don't think automakers will want consumers buying an AV soon because of liability reasons, and we don't see consumers willing to embrace car-sharing in large volume because of the accelerated depreciation.
AVs are not a trend that will go away in a recession. In our view, there's too much brainpower in the technology and auto industries for AVs not to come to fruition. It's also an area where for once, the auto industry and U.S. regulators both want to bring AVs to market. Regulators want to curb fatalities, while automakers want to monetize AV services such as ride-hailing, deliveries for small businesses, and e-commerce. Also, if the auto industry does not change, we think Silicon Valley will steal the industry's pie, so automakers must adapt or die. The AV pie in the United States alone is too big for automakers to ignore. Sales of automobiles in the U.S., assuming 17 million annual sales at about $34,000 a unit on average, is a nearly $600 billion industry. However, if we assume that in 2030, U.S. passengers will travel about 5.1 trillion miles, with consumers paying $0.25 per mile, then mobility is a $1.3 trillion industry. The latter figure assumes no private vehicle ownership, so it's probably too large for 2030, but this possible market size expansion is one reason AVs are drawing so much interest from tech firms and should be a topic investors understand.
For our auto coverage, the fate of private vehicle ownership is the key factor determining what the industry will look like over the next few decades because in some scenarios, the volume contraction of the U.S. fleet is severe. We don't expect private vehicle ownership will go away, especially outside of city centers, but it could change negatively in the latter part of the next decade and beyond in a way that to us suggests automakers taking the lead in AV services, such as
Why AVs Matter Autonomous vehicles will affect so many people because most travel is in passenger vehicles. Road transport, including buses, is over 80% of passenger distances traversed in both the U.S. and in Europe. The U.S., European Union, and China, at say $0.70 per mile, would be a $5.8 trillion market using 2016 passenger miles if everyone used an AV ride-hailing service. The U.S. market alone in 2030 at nearly 5.1 trillion passenger miles traveled, assuming a 1.2% compound annual growth rate from 2016 levels (the average annual change in U.S. passenger miles traveled excluding 2009), at $0.70 a mile is over $3.5 trillion. That figure is over 6 times larger than annual U.S. new light-vehicle sales, which we estimate at about $578 billion, assuming 17 million units sold at roughly $34,000 each currently.
U.S. Driving Habits Set the Stage for AV Transition The 4.3 trillion U.S. passenger miles are, of course, incurred in a vehicle, so there are also vehicle miles traveled. In the roughly 3.2 trillion vehicle miles traveled in the U.S. in 2017, about 78% of vehicle travel happened on roads in urban areas or on interstate highways. When we look at the data in a different grouping, urban roads including urban interstate driving constituted 70% of all miles driven. This fact suggests to us that at a minimum, Level 4 AVs (geofenced autonomous driving) could see wide adoption next decade because there are plenty of companies, including automakers such as GM, mapping cities. AV mapping could then be expanded to suburban areas and used in private vehicle ownership for suburban owners driving town to town within a metro area or from their homes to work, the airport, or downtown. In this Level 4 world, humans could then drive in rural areas off the interstate or drive only when they wanted, perhaps with just a button push or voice command to turn AV capability on or off.
Some investors may fear that AVs will be used 24 hours a day instead of about one hour a day currently, and therefore the U.S. light-vehicle fleet will plummet in size. Data suggests that will not happen unless Americans begin to work night shifts much more than they do now, which we are not expecting, or radically change their behavior so they are awake more overnight and need more mobility during that time. Silicon Valley unicorns such as Uber and Lyft want everyone to share a vehicle; hence services such as UberPool. Lyft cofounder John Zimmer said in an Aug. 23 Wall Street Journalinterview that Lyft is five years into a 50-year vision where it provides a full alternative to owning a vehicle via riding in a Lyft vehicle or on a Lyft bike or scooter. We don't anticipate traffic flow moving to a more balanced distribution throughout a 24-hour day because that would require a huge change in human behavior, and we also don't think everyone is going to want to share a ride all the time to smooth out demand. We say that because the average traffic flow per hour data for the entire seven days of the week shows that peak travel time of 3 p.m. has 304 times the traffic flow of the low travel hour of 3 a.m. and 14 times the 9 p.m.-5 a.m. window. This extra density at peak times creates challenges for AV fleets in routing and in passenger experience. No one is going to want to share an AV if the experience is just like riding the subway or city bus.
Cost/Mile Math Suggests AVs Are Not Coming for Private Ownership Soon We think that for safety and liability reasons, automakers will initially be selling AVs into fleets such as Uber and Lyft as well as making vehicles for their own AV fleets, such as GM's Cruise subsidiary. Costs are another key factor when thinking about the fate of private vehicle ownership. Based on our research, we assume an average cost/mile of $2.50 for a ride-hailing service currently.
Consumers won't give up owning a car purely for financial reasons until AV ride-hailing is noticeably cheaper than the current model. At its November 2017 AV analyst day, GM disclosed expectations that it can price its AV service to customers (currently part of its Cruise division, which may be spun off in the future) at well under $1/mile by 2025 and be profitable, with 20%-30% margins. GM did not specify whether it will be selling AVs to consumers by then, but we think it probably will not due to liability and because these vehicles may only have AV capability in geofenced areas; thus GM will want those vehicles to stay in that area with fleet customers. Even if GM and other AV service providers can offer rides well below $1/mile by 2025, that still could be a material gap above $0.25/mile. Our calculations suggest that a mass switch to AV ride-hailing services instead of owning a vehicle will not happen in the next decade, and we think it may never occur.
Government and Private Sector: Same Destination, Different Map U.S. government regulations, such as the Environmental Protection Agency's emission rules or the National Highway Traffic Safety Administration's corporate average fuel economy rules (known as CAFE), often add complexity to an automaker's product plans because of cost and because rules can conflict with what consumers want. However, for safety and probably job creation reasons, it appears many state lawmakers and the federal government are open to exploring AV development. At the federal level, the AV START Act has not passed the Senate because of Democratic concerns that the bill would give the industry too much freedom relative to safety standards. According to the Alliance of Automobile Manufacturers, the AV START Act will better define federal and state roles of "highly automated vehicles," which we think is critical to avoid a patchwork of regulation where AV laws could vary by state. If the Senate passes the bill, it would need to be reconciled with the SELF DRIVE Act already passed by the House of Representatives in September 2017.
At the state level, according to the National Conference of State Legislatures, 29 states and the District of Columbia have enacted AV-related legislation, with Nevada the first state to authorize use in 2011. Seven states have issued executive orders but enacted no legislation. State acceptance varies, with Nevada enacting legislation requiring a safety driver when AV testing on a highway, to Ohio only issuing an executive order in May 2018 allowing AV testing once companies register with the state. The vehicle must have an operator, but that operator is not required to be in the vehicle. Minnesota's governor signed an executive order in March 2018 that only established an advisory council to study the impact of AVs. We think the best course is federal rules, with states regulating the licensing of occupants/drivers of an AV.
What Does This All Mean? From an auto investing perspective, the biggest issue is the fate of private vehicle ownership. We do not think the costs of relying on an autonomous ride-hailing service will make sense for a long time, if ever. There are also nonmathematical considerations. First, we strongly believe that Americans value their independence, privacy, and freedom. Part of the reason one buys a vehicle is so that it is idle all the time except when you, its owner, need it, even if that means the vehicle is unused about 95% of the day. People do not want to wait for a ride, and we have had too many issues with Uber's mapping causing us to wait, deal with being at an incorrect location, or having to walk to the vehicle instead of the vehicle coming to us to think that Silicon Valley has perfected mapping yet.
Will people want to share a vehicle when they are accustomed to storing some belongings in the trunk or back seat all the time? Will families be willing to wait for an AV ride when two or three kids are crying and the parents just want to go home? We think not. Will parents want to send their children in a vehicle with a stranger who is sharing a ride, or even permit a computer to drive the kids at all? Will you want to share a ride coming home from a bar at 3 a.m. if the extra cost for a private AV is undesirable or not affordable? An AV in this scenario is not as safe as someone riding a crowded New York subway home or a bus filled with people. Along this path, will a wealthy person want to ride side by side with a lower-income person in an AV? With both talking loudly on their phones? Rich and poor together in the back seat of an AV is considerably tighter than riding the subway together. Price discrimination services are likely to arise, but sometimes a wealthy person doesn't want to pay the premium and others may not be able to pay it.