Union of Concerned Scientists
By Don Anair
The roving autonomous vehicles on the streets of San Francisco are one of the frequent reminders on my daily commute that our transportation system is changing. But will self-driving cars be good or bad for climate change?
Imaginations can run wild with “heaven or hell” scenarios of automated cars. Imagine zooming around uncongested roads and highways while passengers attend to their social media, relax with friends, or take in a movie in a clean, electric vehicle. Or, in the darker vision, zombie cars with no passengers are clogging roads and spewing pollution, urban sprawl is given a new life, and marginalized communities continue to lack good transportation options. As this technology comes to market, it will be up to decision makers to set us on the right course with smart policies.
Some researchers have been putting pen to paper to better understand the potential climate risks of self-driving cars (or autonomous or automated vehicles (AVs) as they are otherwise called) as well as their potential climate benefits. This research is providing important insights into the potential for building a modern transportation system that is less polluting, less congested, more equitable and more efficient than what we have today. It also highlights the significant risks of inaction and the difficulty of achieving the best outcomes.
3 Revolutions and a Multi-Modal Future: Autonomous, Electric, and Sharing Rides
Let’s start with the positive vision first. Self-driving car technologies are paired with electric vehicles, which we’ve shown have lower carbon emissions no matter where you live in the U.S. In addition, AVs usher in a new wave of transportation services—think Uber and Lyft 2.0—where rides are more convenient than individual vehicle ownership and are cost-competitive. This leads to a reduction in personal car ownership, since not owning a car is now a more viable, cheaper option for households. Reduced car-ownership alone doesn’t solve the problem, but when paired with increased access to mobility options like shared bikes, scooters, and efficient mass transit, individuals now choose from a variety of options for each trip, rather than always defaulting to the car formerly parked in their driveway. Sharing or pooling of rides is seamless and offers a lower-cost option, access to faster moving car-pool lanes and lower tolls, while reducing the number of cars on the road. This ideal future of clean, equitable, and accessible mobility is one of autonomous, electric, and pooled car trips combined with urban design and infrastructure that supports walking, scooters, bikes, and mass transit, and pricing signals that steer choices towards the cleanest, most efficient modes of travel.
What happens to climate emissions in this future? Researchers at University of California Davis and Institute for Transportation & Development Policy examined a future scenario where AVs are incorporated into a highly shared, multimodal, and electric urban transportation system. They found, globally, urban transportation pollution could be reduced by 80 percent by 2050 and massive increases in congestion could be avoided, with vehicle miles traveled actually declining by 25 percent instead of increasing by 50 percent in the business as usual case (see figure).
This scenario of a future transportation system meets the travel demands of a growing population while driving down climate emissions. And it requires coordinated policies to work, including compact development as well as policies that make the lowest emission and most efficient modes of transport the most attractive. But what if that’s not what happens? What if we don’t make the decisions necessary to support the future described above, and instead take a hands-off approach to AV deployment?
The nightmare AV future: More vehicle miles, more congestion, more pollution, less equity
As wonderful as the vision of “three revolutions” is, it would be foolish to think that this vision of the future is likely—or even possible—without a lot of work. Here are a few ways that things could go wrong.
AVs could dramatically increase driving
If AVs primarily enable increased single occupancy vehicle trips, we are in trouble. One widely-cited study looked at a wide range of impacts AVs could have on energy consumption, travel and carbon emissions. And there are many factors (see figure). Everything from the energy savings of robot eco-driving to energy and travel increases from newly empowered individuals who previously did not have the ability to drive their own vehicle. There are several potential impacts on both sides of the ledger, but the biggest potential increase in energy use (and by association, emissions) comes from a behavioral response to AVs. If driving can now be productive time, longer commutes, for example, may not be the burden they once were. This is one way in which AVs could reduce the time-cost of driving (see “travel cost reduction” results in the figure) and increase overall vehicle travel – as much as 60% according to the study. Recent modeling of possible AV deployment in the Washington, D.C. metro region showed similar results, estimating that vehicle miles traveled could increase 46-66% with the introduction of self-driving cars.
So will people really drive that much more? Some researchers did an experiment to see what would happen to a household’s vehicle travel if they had access to a vehicle and a driver for a week – mimicking life with a self-driving car. Not surprisingly, most households used the vehicle more often (83% average increase in miles traveled), and even sent the car and driver out on errands (21% of the increase was zero-occupancy). While there were only 13 participants in the study, which limits the generalization of the findings, the experiment does illustrate the potential behavioral shifts when a vehicle that can drive itself is introduced into a household. Why not send the car to pick-up your dry cleaning or take that trip to Aunt Esmerelda’s you’ve been putting off?
AVs could increase congestion and undermine transit, instead of complementing it
Pooling rides is essential to making AVs deliver on their potential to be clean, equitable and efficient. Pooling rides for people with similar origins and destinations can deliver more passenger trips from fewer vehicle trips, which is key to making efficient use of vehicles (reducing pollution per trip) and roads (reducing congestion per trip). But while pooled AVs could help increase the average occupancy of cars, they could also undermine our most important current source of pooling, mass transit. A car with 2-3 people sharing a ride is an improvement over each person driving alone, but it is a lot more vehicles, pollution and congestion than 30 people in a bus, or several hundred in a subway or train.
Based on the current evidence, especially in larger cities where mass transit is especially important, ride-haling is pulling more people from modes like transit, walking and biking than it is pooling passengers who would otherwise drive alone. This mode shift, along with additional trips that that wouldn’t have been made in the absence of ride-hailing options, is leading to increases in congestion and increased vehicle miles traveled. (See research by Clewlow & Mishra, Schaller, and University of Kentucky) Moreover, reduced ridership on mass transit hurts the economics of these critical systems as they lose fare revenue. Adding AVs to ride-hailing fleets could drive down ride costs and exacerbate the changes in vehicle travel and transit impacts we are already seeing.
Roads snarled in congestion are not a good outcome for anyone, including companies that want to use these roads to sell people rides, pooled or otherwise. So, new rules and incentives will be needed to efficiently manage transportation networks as private companies operate what are in effect private transit systems with occupancy sometimes higher than today’s cars but most often lower than today’s mass transit. Policy-makers will need to prioritize the movement of people over vehicles with policies that favor higher occupancy trips and modes. These could take the form of preferential pricing, access to restricted lanes and ensuring that the financial model of mass-transit adapts along the way
If we don’t succeed in ensuring rides are largely pooled in both cars and in mass transit modes like rail and subway, not only will congestion get worse, but we will fail to reduce climate emissions to safe levels as electrifying our transportation system is simply not enough. In the UC Davis/ITDP study, a “2 Revolution” scenario with AVs and widespread electrification but WITHOUT significant pooling of trips resulted in emissions reductions globally in 2050 by only 45% – far less than needed to stabilize our climate.
AVs could exacerbate or perpetuate inequities in our current transportation system
Our current car-ownership-based transportation system does not serve all communities in an equitable way. Lower income households spend a larger share of their income on transportation than wealthier households. Those who cannot afford a car, or are too old or young to drive, or have physical handicaps to driving, have to rely on a transit system that often doesn’t meet their needs.
AVs could improve mobility for communities historically underserved by our current transportation system – if the technology enables greater access to affordable, accessible and reliable transportation. If, however, AV technology is primarily relegated to private car ownership and leads to increased congestion or undermines public transit, as described above, the current inequities will be exacerbated.
A new report by the Greenlining Institute describes in more details the health, economic and mobility risks of AVs for marginalized groups like people of color, the poor, the elderly, and those with disabilities, and offers a list of recommendations to policymakers for ensuring the rollout of AVs leads to greater mobility options for all. UCS will also be releasing a report soon with results from an analysis of the Washington DC metro area and how the rollout of AVs in that region could impact transportation equity. This research is important for informing the policies necessary to maximize the benefits of self-driving technology.
Now’s the time to get on the right path
Research is providing some helpful insights on understanding the potential role of AVs in a transportation system that cuts climate emissions and improves mobility. It also offers a cautionary tale of the potential for AVs to dramatically increase emissions and exacerbate congestion if decision makers are not proactive and thoughtful about putting in place the policies that will lead us to the best outcomes.
We are starting to see some positive action on this front. In California, legislation (SB1014) signed into law last year requires state agencies to develop standards to ensure ride-hailing companies are moving towards greater shared, zero-emission trips. Since AVs are likely to be rolled out in ride-hailing services, these rules will affect AV deployment. But that’s only a drop in the bucket. Developing effective public policy to ensure AVs deliver climate and transportation system benefits requires shared goals, effective interagency coordination, and development and implementation of effective policy at different levels of government. In California, UCS is sponsoring legislation with CALSTART (SB59 authored by Senator Ben Allen) that would get the ball rolling at the state level and ensure proactive policies can be deployed as AV technology is hitting the street.
Smart policies are critical for ensuring self-driving car technology ushers in a new era of clean, affordable, and efficient transportation rather than the zombie car apocalypse. AVs may be able to drive themselves, but it is up to us to steer them in the right direction.