An NBER Conference on the Economics of Autonomous and Electric Vehicles took place June 6-7 in Stanford University. Research Associates Susan Athey of Stanford and Ryan Kellogg of University of Chicago, and Jing Li of MIT organized the meeting, which was sponsored by the Alfred P. Sloan Foundation. These topics were discussed:
Resolving Braess's Paradox through Information Design:
In the era of the Internet of Things (IoT), devices are all connected, which makes it possible to gather and share a vast amount of real-time data. Besides technology advancements, the way to distribute information is also critical to increase social benefits. Liu and Whinston look specifically into the information-based routing problem for heterogeneous autonomous vehicles. To address the issue, they synthesize theoretical economics and artificial intelligence (AI). In a fully autonomous transportation system, based on the idea of Bayesian persuasion introduced by Kamenica and Gentzkow (2011), its extension by Kolotilin et al. (2016) and the unified information design framework by Bergemann and Morris (2017), the researchers propose a method to design the information structure conveyed to a vehicle based on its private preference, strategic interactions with other vehicles, expressway tolling, and other traffic conditions. In a simple road system, they present both a multi-vehicle model and a multi-period model. They also build up a framework to deal with urban traffic. To implement the method they propose, Liu and Whinston introduce the idea of creating a hardware engine to accelerate the calculation and demonstrate it by a flow chart.
GHG Implications of an Autonomous Future
As has been well documented, the transportation sector is currently undergoing a profound transformation in the way that vehicles are powered, owned, and operated (Center for Automotive Research, 2016; Fagnant and Kockelman 2015; Sperling, 2017; Lovejoy, Handy, and Boarnet, 2013; Arbib and Seba, 2017). The complex interdependence of these transformations, as well as the fact that Hatch and Gormane are studying them concomitantly with their development, makes them particularly difficult to predict. Recent studies have shown preliminary evidence that ride-hailing has shifted travel miles dramatically away from traditional forms of transport such as public bus and rail service, biking, and walking, and towards ride hailing services, resulting in a significant increase in VMT (Clewlow and Mishra, 2017). At the same time, fleet ownership business models may have a significant impact on manufacturing emissions and vehicle efficiency. Building upon the model framework developed by Fox-Penner, Gorman, and Hatch, the researchers consider the greenhouse gas implications of various light-duty transportation business model adoption scenarios, including the effects on manufacturing inputs, technology improvements, and total vehicle miles traveled. Hatch and Gorman's goal is to build a more realistic and research based estimate of future transportation scenarios and their impact on personal vehicle emissions.
The Potential Distributional Impacts
Minimizing Fleet Emissions through
The bulk of federal- and state-level subsidy programs for electric vehicles (EV) offer these funds at a flat rate. If the goal of these initiatives is to minimize vehicle emissions, the flat subsidy design ignores several issues a more efficient program should feature. The first is that funds should target marginal consumers whose likelihood to purchase EVs increase most with the vehicle subsidy. The second that the subsidies are directed toward consumers most likely to impact emissions by replacing less energy efficient vehicles. To address these two points, Seo and Shapiro propose a subsidy design which maximizes the expected reduction of emissions per dollar spent by targeting marginal consumers with the highest impact on vehicle emissions. Their analysis exploits spatial discontinuities in California's local subsidy programs and utilizes zip code-level vehicle purchase data to estimate demand with rich heterogeneity across consumers. They then characterize the marginal consumers for EVs along key demographics, such as income and vehicle ownership. Using vehicle stock and replacement behavior, Seo and Shapiro then sort these consumers by how much their vehicle replacement would reduce emissions.
Vehicle Depreciation and Survival
Technological Frontiers and Challenges
Carpooling and the Economics of Self-driving Cars
Ostrovsky and Schwarz study the interplay between autonomous transportation, carpooling, and road pricing. They discuss how improvements in these technologies, and interactions among them, will affect transportation markets. Their main results show how to achieve socially efficient outcomes in such markets, taking into account the costs of driving, road capacity, and commuter preferences. An important component of the efficient outcome is the socially optimal matching of carpooling riders. Ostrovsky and Schwarz's approach shows how to set road prices and how to share the costs of driving and tolls among carpooling riders in a way that implements the efficient outcome.
The Electric Vehicle Transition and
Holland, Mansur, Muller, and Yates analyze the transition to electric vehicles and recent policy proposals to ban gasoline vehicles. Their model captures declining electric vehicle damages; declining electric vehicle production costs due to exogenous changes or to learning by doing; stock effects; and the introduction of complementary infrastructure such as charging stations. The researchers derive conditions under which it is socially optimal to ban gasoline vehicle production in the long run. Holland, Mansur, Muller, and Yates derive two classes of solutions. In one, it is optimal to ban gasoline vehicle production before beginning production of electric vehicles. This solution obtains if electric vehicles are perfect substitutes for gasoline vehicles. In the other solution, it is only optimal to ban gasoline vehicle production after beginning the production of electric vehicles. Simulation results show that the optimal time to ban gasoline vehicles depends critically on the parameters that describe preferences for the two types of vehicles.
Ridesharing, Spatial Frictions,
With the pervasiveness of mobile technology and location-based computing, new forms of smart urban transportation, such as Uber and Lyft, have become increasingly popular. These new ridesharing (or e-hailing) platforms can influence individuals' movement frictions, in turn influencing local consumption patterns and the economic performance of local businesses. To gain insights about the future impact of urban transportation changes, Zhang and Li analyze individuals' urban consumption patterns before and after their adoption and usage of the ridesharing services. Their study is validated using a novel, anonymized panel dataset of finegrained individual credit card and debit card transactions from January 2012 through May 2016 from a large U.S. bank, including from over 10K observed adopters of ridesharing with 7M transactions. Based on revealed preferences, the researchers hypothesize that those who choose to use ridesharing more are those whose mobility frictions are affected more by the availability of ridesharing, and vice versa. Their findings demonstrate a significant positive impact from the usage of ridesharing services on individuals' consumption frequency, as well as the spatial diversity of their spending. Zhang and Li's results also indicate strong heterogeneity in such effect. The effect becomes significantly stronger in increasing restaurant and bar transactions, and also stronger among younger customers and those who spend less before adopting the ridesharing services. Lastly, the reseachers see evidence of the heterogeneous impact of spatial frictions between different neighborhoods of a major city.
Congestion and Incentives in the Age of Driverless Cars
GPS systems and Autonomous Vehicles (AVs) will likely open the way to forms of traffic coordination, or centralization. Boffa, Fedele, and Iozzi analyze the welfare effects of moving from an environment with atomistic drivers to one in which few companies will manage the traffic. Differently than what happens with atomistic drivers, such companies or organizations will have an incentive to consider the congestion externality imposed by their vehicles on the other vehicles they dispatch. The researchers analyze both a setting with no road taxes, to reflect their limited application and the popular opposition to them, as well as a setting with road taxes. They find that, without road taxes, the emergence of a small company supplying a small fraction of the travelers (while the others remain atomistic) increases (decreases) welfare if and only if the congestion problem was (was not) sufficiently severe in the first place. With road taxes, the researchers find that, while congestion charges are optimal when all travelers are atomistic, the structure of the taxes differs markedly with a company that supplies a mass of customers. Restoring first best, in this case, may require subsidizing the company - something likely to be politically very unappealing.
Impact of Vehicle Automation on Electric Vehicle
Mersky and Samaras present a method to characterize the impact of privately-owned autonomous electric vehicles on electric vehicle charger placement, distribution, utilization, and power demand. Using Seattle, WA as a case study, a least total cost optimization for charging station owner and driver costs is conducted for vehicle automation levels 0-3, 4, and 5. Moving from levels 0-3 to level 4 and level 5 automation reduces the peak electrical load for EV charging by approximately 31% and 68%, respectively. Moving from levels 0-3 to level 4 automation decreased the optimal number of chargers by 65%, lowered total cost by 46%. Moving from levels 0-3 automation to level 5 automation decreased the optimal number of chargers by 84% and total costs by 69%. Additional vehicle miles traveled and operating costs incurred by drivers for drop off and pick up were estimated with level 5 automation. The results suggest that highly automated vehicle technology used in privately-owned electric vehicles could reduce the cost of deployment for recharging infrastructure and reduce peak electrical demand associated with recharging.
Market Expanding or Market Stealing?
The recent rise of dockless bike-sharing is dominated by two firms in China: one started first in 82 cities, 59 of which were subsequently entered by the second firm. Using these variations, Cao, Jin, Weng, and Zhou study how the entrant affects the incumbent's market performance. To our surprise, the entry expands the market for the incumbent. Not only does the entry boost its total number of trips and encourage more bike investment, it but also allows the incumbent to achieve higher revenue per trip, improve bike utilization rate, and form a wider and more evenly distributed network. The market expansion effect on new users dominates a significant market-stealing effect on the incumbent's old users. These findings, together with a theoretical model that highlights consumer search and network effects, suggest that a market with positive network effects is not necessarily winner-takes-all, especially when users multi-home across compatible networks.