other car on the road. In larger simulations, once autonomous vehicles reach a level of 5 to 10 percent of all cars on the road, they can manage localized traffic in complex environments—such as when eight lanes of traffic merge into two or at an extremely busy intersection. An autonomous vehicle carefully increases the space between itself and the car ahead, anticipating that it will abruptly slow down at some point down the road. However, because the self-driving vehicle has created a buffer, it can keep driving at the same pace—avoiding the cascade effect that its own braking would have on the cars behind it. In time, a new equilibrium is reached where all the cars are moving at a steady pace.
Daziano, Sarrias, and Leard (2017) takes a different perspective on travelers’ benefits from autonomous vehicles and finds that the average household’s willingness to pay for those vehicles greatly exceeds the additional cost of the technology. To be sure, our rough calculation and the preceding studies should be qualified because the actual benefits of autonomous vehicles will depend on whether the government reforms policies that would enhance their driving environment and performance. They will also be affected by the extent to which travelers reduce the capital costs of vehicle ownership by sharing vehicles through transportation subscription services.
The Role of the Government
Applying autonomous-vehicle prototypes to multiple geographies is difficult because cities have different road infrastructures—from signs to road paint—and different norms that drivers follow. In addition, no public benchmarks exist for how autonomous vehicles should perform in a given scenario to meet a standard of safety. Accordingly, all levels of government have a critical role to play in the public’s adoption of autonomous vehicles because the public highway infrastructure is provided by local, state, and federal governments and because the federal government is generally responsible for regulating motor-vehicle safety.
Other major innovations have also relied on infrastructure provided by the government. For example, cell phones and eventually smartphones relied on the government-provided electromagnetic spectrum to enable communication through cellular technology. Similarly, autonomous vehicles must depend on the public road system to transport passengers and freight by vehicles that drive themselves. Because inefficient infrastructure policies have contributed to the large social costs of nonautonomous vehicle travel, it is important for policymakers to reform highway-infrastructure pricing, investment, and production policies to reduce those costs for autonomous-vehicle travel. In addition, local and state governments will need to make investments to upgrade the physical characteristics and communication capabilities of their infrastructure to facilitate autonomous-vehicle travel.
In 2018 Congress drafted but failed to pass important autonomous-vehicle legislation that would have clarified and expedited vehicle testing and possibly vehicle adoption. The House passed self-driving legislation, called the Self-Drive Act, but the Senate failed to pass its version, the AV Start Act. Five Democratic senators blocked floor consideration and a final vote on the bill, and Senate Majority Leader Mitch McConnell chose not to use up limited floor time by bringing a cloture motion to defeat the holds. As this book goes to press, Congress is considering but has failed to pass a revision of its autonomous vehicle legislation.
The National Highway Traffic Safety Administration (NHTSA) could have initiated motor-vehicle-safety rulemakings without additional federal legislation. The only step forward that NHTSA has taken was to approve in February 2020 its first exemption from safety standards for a self-driving car. The exemption is limited to Nuro’s R2, which is not designed to carry people but instead to deliver groceries or pizza and to have a top speed of 25 miles per hour.
The purpose of the congressional acts was to mandate a conflicts audit and recommendations for future rulemaking amendments and additions and to greatly expand exempted annual vehicle volumes while future amendments and additions are being developed over the next decade. Given that NHTSA has not made any major rulemakings, it is important for Congress to pass autonomous-vehicle legislation to ensure that the agency takes appropriate actions in a timely fashion, to expedite regulatory development pertaining to vehicle design and performance, and to provide regulatory relief where appropriate. Specifically, the legislation would allow the secretary of transportation to provide regulatory exemptions for a maximum number of vehicles per automaker per year, thereby enabling automakers to test their vehicles without any delay in advance of NHTSA’s setting the final standards through the Federal Motor Vehicle Safety Standards at some future date.12
Finally, new legislation would affirm federal responsibility for the safety of autonomous vehicles and would prevent states from instituting their own vehicle regulations beyond their normal responsibilities of licensing and registration. Thus automakers and technology companies would not have to navigate through a patchwork of different regulations to test their vehicles and could work with regulators both to determine the safety goals that their vehicles would be expected to achieve and to identify how those goals would be met.13
As is the case with nonautonomous vehicles, individual automakers would be required to ensure their compliance with the safety standards by self-certification. After NHTSA sets the final standards, makers of autonomous vehicles would be free to sell their vehicles to the public, and consumers would begin their transition from driving nonautonomous vehicles to traveling in autonomous vehicles. The public’s adoption of autonomous vehicles would therefore occur after NHTSA had set its final design and performance standards, which could take several years, during which time agency officials observed automakers’ testing of autonomous vehicles and improvements in the technology. Widespread adoption would only occur once the majority of travelers had purchased or arranged to share AVs and no longer use nonautonomous vehicles, which could take several more years. In sum, it could be at least fifteen to twenty years before the public’s adoption of autonomous vehicles exceeds 50 percent, and it could be another ten years before autonomous vehicles are completely adopted and used throughout the country.14 Of course, positive or negative government policy or nonpolicy shocks to the adoption process could alter that timetable.
In the absence of federal legislation that has affirmed the federal government’s role in the testing and adoption process, some cities and states have started to pass their own regulations to allow testing of autonomous vehicles within their borders. The number of cities conducting or planning to conduct a test is steadily growing.15 However, federal legislation must be passed to allow the formal adoption process to begin.16
Notwithstanding the delays in instituting federal regulations, autonomous-vehicle providers are already competing intensely in technology so that they can deploy their vehicles as soon as the federal legislation is passed, increasing the likelihood that they will survive the anticipated shakeout when final regulations are in place and consumers start to adopt autonomous vehicles. None of the autonomous-vehicle companies want to suffer the fate of Nokia and Blackberry, driven out of the market by a superior technology, Apple’s iPhone, which was promptly introduced when the infrastructure, telecommunications bandwidth, could accommodate it and when new entry and intense competition ensued.
Certain countries, such as China, may be able to expedite public adoption of autonomous vehicles because its citizens are seemingly more trusting of the technology than are those of other countries and because the government can more easily overcome political obstacles when it wants to initiate regulatory change. The United States and Japan, among others, are also sensitive to global competition, and each wants to be the first country where autonomous vehicles are widely used. However, some industry observers (for example, Wenderoth 2018) predict that Baidu, a Chinese company, will capture the headlines as the first company to sell autonomous vehicles to its nation’s travelers, rather than Waymo, GM, or Toyota. If so, the United States should be more concerned with the large welfare costs that residents have incurred because of unnecessary regulatory delays to autonomous-vehicle adoption than with the blow to its national pride.
Part 2
Potential Effects of Autonomous Vehicles
