The era of the driverless car seems closer than ever. A few weeks ago, Singapore announced its plans to roll out driverless pods by the end of 2016. Germany is working on autonomous cars that will transport passengers to train stations, as well as automating the trains themselves. And here in the United States, a recent Wall Street Journal article showcased a new plan from General Motors and Lyft to begin operating a pilot program of autonomous taxis within the next year (the lucky city has yet to be determined).
As the WSJ notes: “one of the top hurdles to their success is navigating a patchwork of regulations that govern the use of autonomous vehicles and liabilities.” The Mercatus Center’s Adam Thierer and I discussed many of those regulatory hurdles in a 2014 paper. We also discussed many of the clear benefits that would emerge from autonomous vehicles at a high level of market penetration—most notably, abating many of the 33,000 annual fatalities that occur on American roadways due to human-operated error. However, there are many other benefits we didn’t touch upon that are worth considering.
The potential environmental impacts of autonomous vehicles are not well understood. This uncertainty is in large part driven by an inadequate ability to account for how they will be utilized when they enter the market. Will they function as a service industry, rather than a capital good? Will that industry be predicated on a pooled or shared autonomous vehicle (P/SAV) model, or driven by individual consumption popularized by Uber? How will travel patterns change as a result? Will people rack up more vehicle miles traveled (VMT)? Will more cars be operating more frequently? What will be the proportion of electric to internal combustion engine (ICE) vehicles? These and many other questions abound, which make a clear determination of the associated benefits difficult to quantify.
Despite the inherent difficulties, the McKinsey Global Institute has suggested that CO2 emissions “could be reduced by as much as 300 million tons per year,” which would be the equivalent to a 50 percent reduction of CO2 currently emitted as a result of commercial aviation (download the full report in PDF form here). A report from the University of Texas-Austin predicted that for every SAV integrated on Austin’s roadway, 9 conventional vehicles could be replaced, resulting in net environmental impacts, even when accounting for a potential 8 percent increase in VMT.
Additional environmental benefits would accrue from fewer “cold starts,” dynamic ridesharing, and a higher maintenance turnover rate. (That is, because the vehicles would be travelling more often, maintenance would need to occur more regularly than with conventional human-operated vehicles, which “will allow fleet operators to consistently provide SAVs with the latest sensors, actuation controls, and other automation hardware, which tend to be much more difficult to provide than simple SAV system firmware and software updates.”)
Additionally, it seems likely that autonomous vehicles will be incorporating electric batteries as power sources over conventional internal combustion engine (ICE) vehicles. In 10-15 years, possibly sooner, it’s likely that we’ll see electric batteries arrive at a price point well below $120 kWh—far below the $150-160 kWh price point necessary to make electric car batteries price competitive with ICEs. Batteries are only getting cheaper, and in the long term, traditional fuels will only get more expensive. Even when accounting for production supply chain emissions, the environmental benefits of battery-powered electric vehicles outweigh traditional ICEs with “excess manufacturing emissions … offset within 6 to 16 months of average driving.” And reductions in electric vehicle emissions will only continue falling in the years ahead.
Assuming those trends continue—and there’s little evidence to suggest technological development and advances in chemical engineering are set to stall anytime soon—it seems likely to conclude that when economies of scale permit higher adoption rates, most autonomous vehicles will run on electric engines, not ICEs.
Congestion and Fuel Efficiency Cost Savings
Approximately 25 percent of traffic congestion is attributable to traffic incidents, particularly vehicular crashes largely resulting from human-operated error. Here again, autonomous ridesharing could have profound effects. Examining the effects of congestion in Boston, an MIT study found that a 50 percent ridesharing adoption rate could potentially result in a 37 percent reduction in congested travel time, as well as a 19 percent reduction in total vehicles on the roads.
In an October 2013 report from the Eno Center for Transportation, congestion and fuel economy benefits were determined to be the primary drivers of economic savings from autonomous vehicles. Annually, almost 3 billion hours of travel time could be saved as a result of less congested roadways. If, but more likely when, cooperative adaptive cruise control (CACC) achieves a 90 percent market penetration, the effective capacity of road lanes would increase by 80 percent. The operational efficiency of intersections could also be increased by reducing intersection delays, though these gains will likely require high levels of autonomous vehicle market penetration. Fuel savings are likely to be among the leading drivers (pun intended) of economic gains from adopting this technology.
Sensor suites respond near-instantaneously to the actions of proximate vehicles, resulting in diminished fuel waste that occurs due to rapid acceleration, speeding, and variations in speed. Additionally, given the rapid response of computerized sensors, autonomous vehicles would be capable of traveling more closely together. This would result in reduced air resistance and improved fuel efficiency by as much as 20 percent, especially if the cars were platooned—that is, connected via vehicle-to-vehicle communication signals that help optimize speed and minimize the distance between vehicles.
But these gains are low-end, conservative assessments of the potential for autonomous vehicles. In fact, if we were to assume a 90 percent adoption rate, American consumers would save an estimated 724 million gallons of fuel every year. With the current national average price of gasoline hovering around $2.20 as of this writing, that amounts to almost $1.6 billion in savings.
Of course, many of these savings are predicated on assumptions regarding the general makeup of any autonomous vehicle fleet. The need for diversity in the fleet will be a key, and often unmentioned, necessity in maximizing many of the potential benefits in these areas. For example, a single individual calling an autonomous vehicle for a ride to work along a route with minimal inclines would require a different vehicle than a family of four taking a trip across rolling hills. Optimizing the fuel efficiency gains for the former scenario requires a different vehicle than for the latter, and optimizing algorithms for matching vehicles to particular needs will be necessary to realize gains in fuel efficiency.
Transit for Low-Income Urbanites and Reduced Land Usage
Improved mobility for low-income urban residents is also a benefit that has received underwhelming investigation. Residents in poorer neighborhoods have long suffered from limited access to optimal transportation networks than their wealthier counterparts. Autonomous ride sharing fleets could go a long way towards giving low-income urbanites greater access to a larger portion of the urban landscape, and therefore increasing the ease of accessing a larger job market.
These benefits are far more likely for urban residents than rural or suburban, given the amount of VMT. Assuming the cost of an SAV service model were to range between the prices of carsharing ($0.60-1.00 per mile) and human-driven taxi ($2.00-3.00 per mile) models, SAVs could be a potentially attractive alternative for those individuals travelling less than 5,000 miles annually (as opposed to purchasing a vehicle of their own). Alternatively, some estimates place the potential per-mile cost of SAVs to run as low as $0.12 by 2040. Examining the potential cost-savings to consumers in Manhattan, one study found the potential for reducing average trip costs from $7.80 to just $1 as a result of excising the costliest element of the taxi market: human labor.
Increased independent mobility for low-income individuals would also minimize the need for substantial investment in public transportation systems. Though such transportation networks would still be a mainstay of the urban cityscape, supplementing their operation with SAVs could serve as an effective means of alleviating congestion during peak travel times.
The amount of land needed for parking would also likely diminish with SAVs. This should come as no surprise, given that the average car spends approximately 95 percent of its lifetime parked and immobile. Given that upwards of a third of all land in urban centers is occupied by surface parking lots and garages, the incorporation of SAVs could free up land for commercial and/or residential development, driving down the cost of rent and further contributing to ameliorating the plight of low-income urbanites. If we account for the total social cost of parking infrastructure—that is, costs associated with raw materials extraction and processing for construction, the actual construction of lots and garages, and operating and maintenance expenses—annual costs range between $4-20 billion annually.
Couple these types of developments with advances in robotics and other forms of automation, as well as advances in additive manufacturing, and city-based construction costs could plummet by 20 percent, further inducing economic gains.
When portending the future, we tend to view it through rose-colored glasses. After all, it’s entirely possible that autonomous vehicles will end up being a major commercial market for private consumption rather than service-based ridesharing models. If the future bends in that direction, many of the aforementioned benefits will likely fail to materialize. However, with each subsequent Uber ride, people seem more and more willing to ditch the old allure of traditional vehicle ownership with the ease and convenience of ridesharing apps, especially the younger generation.
In the near term, these trends are likely to be limited to urban centers, but those are the areas where the benefits of autonomous ridesharing will be most pronounced. Rural areas will likely, unfortunately, lag far behind their urban counterparts in adoption rates for SAVs. But we shouldn’t let that blind us to the possibilities of an autonomous future. By 2025, the total value gains from autonomous vehicles could be as high as 90 percent fewer accidents, 20-40 percent gains in fuel efficiency, and $8 per hour in value of travel time saved. The total sum of potential economic impacts could be as high as $2 trillion. Whether those estimates are high or low, the benefits of this future will, on net, be a boon to us all.