The Future of Transportation Disruption and How Public Agencies Can Respond
This is the second and final article in this series. Read the first article here.
Our modeling results confirmed the expectations noted above, but the magnitude of the effects may be surprising. Figures 1 and 2 show the range of effects captured by the models for vehicle travel and transit use. Each dot in the chart represents the results from an individual model and the tests included both private ownership of AVs similar to automobile ownership today plus a scenario where 50 percent of drive-alone (i.e., single-occupant) trips were shifted to shared vehicles (i.e., carpools).
This ‘shared’ scenario is intended to capture the potential influence of government policies or regulations to encourage much higher levels of ride sharing than have traditionally occurred with privately owned vehicles. The models were not capable (without substantial modification) of capturing the zero-occupant vehicle trips that would occur as AVs travel between different passengers or to final parking locations when not in use. Also, long-term land use changes were not accounted for in these model applications. So, higher levels of VMT are possible.
While the results are best viewed for the potential range of effect. The averages are also provided below.
- Vehicle trips increase by an average of 20% without any shared-use regulation. That increase is virtually eliminated (on average) with 50% of the AVs required to be shared rides.
- VMT increases by an average of 31% without any shared-use regulation. That increase is halved (on average) with 50% of the AVs required to be shared rides.
- Transit trips decline by an average of 29% without any shared-use regulation, which grows to 35% with 50% of the AVs required to be shared rides.
- On average, bus and transit trips less than 5 miles decrease more than rail and transit trips greater than 5 miles.
In comparison to other research we see that these results may be highly relevant. For example, a unique experiment involving the provision of 60 hours of free chauffer service for one week (2) and showed a VMT increase of 83 percent for those participating. While, this experiment was conducted using a small sample of 13 test subjects from the San Francisco Bay Area it underscores the significance and importance of our modeling.
While the results may be due to different model strengths and weaknesses rather than real-world variations in effects. They did not capture all induced growth and induced vehicle travel effects. The model tests themselves were designed as ‘stress tests’ to better understand potential effects and level of sensitivity to help inform future research and analysis.
What Can Policy Makers Do Given These Future Outcomes?
These modeling results are important at framing future behavior and transportation trends, but they also underscore the policy context. The actions of government can have a dramatic effect on these outcomes. Policy or regulatory responses can change costs of a ride or the number of people in a vehicle. Governments can use policy and regulation to balance the desires of private companies with the public good.
With that context in mind, we provide a brief list of potential policy and regulatory responses designed to offset the effects revealed by the modeling tests. In general, the responses include: increasing public transit competitiveness; increasing the occupancy of new mobility vehicles, decreasing their size, and increasing the cost of zero-or-low-occupancy vehicle travel; and using land use policy.
Part of the explanation for the decrease in transit ridership is that transit travel times are much slower than automobile travel especially if delivered in a TNC or MAAS platform door-to-door. Today, TNC passengers have very short wait times, often less than 5 minutes, and in-vehicle travel times that are similar to using a private auto. AVs could improve upon the wait time and possibly the in-vehicle travel time due to capacity increases. AVs will also reduce the cost of vehicle travel. In response, transit travel experiences and travel times need to improve to remain competitive as outlined below.
- Increasing frequency of service – Frequency directly influence wait times and provides flexibility to system users to come and go from destinations without having to worry about schedules.
- Providing transit-only lanes – Similar to operational hours, in-vehicle travel time on buses needs to be faster to compete with vehicle travel. Transit-only lanes (during peak periods) improve roadway space efficiency and utilization and would lower current in-vehicle travel times.
- Automating transit service – Buses on fixed routes are one of the first opportunities for autonomous vehicle use. Fixed routes are easier to navigate than an open network and the switch to autonomous operations reduces labor costs. Savings could be redirected to expanding core services especially rail service, which would minimize the impact of driver reductions by increasing other operational and maintenance jobs. Savings could also be translated into reduced fares, with the possibility of offering free transit service that could stimulate a virtuous cycle of attracting more ridership and reducing cost per rider such that service could be expanded to attract even more riders. If automated operation is combined with technology for matching riders and vehicles, then autonomous rapid transit (ART) service could be offered.This type of service would operate in transit only lanes but have the benefit of matching riders with common destinations together in transit vehicles. This type of matching would allow ART vehicles to skip some stations once the vehicle is full thus improving in-vehicle travel times compared to conventional transit.
- Better match or ‘right size’ transit demand to type of service – TNCs, MAAS, and AVs offer expanded options for demand-responsive and crowd-sourced transit in low- to medium-density areas whether service is provided by the public or private sector. Private sector TNC platforms benefit from costs only being incurred when a trip is made. Public agencies could benefit if allowed to operate this type of on-demand door-to-door service using a similar platform or by contracting for this type of cost-effective service when traditional fixed-route bus productivity would otherwise be low (i.e., less than 10 riders per hour). Being more cost-effective could also allow for extended operating hours that are necessary for transit to provide reliable all-day travel.
Increasing Occupancy or Decreasing Size
In terms of AVs, absent government regulations, initial implementation will likely occur through TNC or MAAS platforms (although companies like Tesla may offer AV technology through a traditional private ownership model). While AVs offer potential benefits such as reducing collisions they also make vehicle travel more attractive. Increases in vehicle use could exacerbate current problems associated with congestion and emissions especially if vehicle sizes remain large and occupancy levels remain low. The following actions are intended to minimize adverse effects of greater vehicle use.
- Require AVs to be electric – Using electric power generation would minimize the emissions associated with AV travel.
- Support small or micro-sized AVs for personal use –The image below shows how vehicle size influences intersection delay and fuel consumption. Today’s large vehicle sizes (combined with low occupancies) consume substantial physical space, capacity, and green time at signalized intersections. Reducing vehicle sizes improves network performance.
- Manage or price AV travel to encourage high occupancy levels – Various studies of AV effects emphasize that the only way to prevent substantial increases in vehicle use (i.e., VMT) is require AVs to operate as taxis carrying multiple passengers. Building pricing into AV use early can help shift ground transportation towards more efficient travel outcomes and to partially offset transportation revenue losses from parking and citations. Instead of peak period travel demand routinely overwhelming available roadway supply, demand-responsive pricing of AVs especially in TNC or MAAS platforms could help manage fleet sizes and roadway space utilization. This policy response is not simple though and would require addressing numerous issues typically raised for any U.S. road pricing proposal. Notably, equity of any change given current system is publicly owned and perceived as ‘free’ by users, absence of an existing market (creating prices does not necessarily create a market), use of revenues to ensure efficient outcomes, limited ability to transfer payments from those willing to pay for travel to those willing to forgo travel, and whether revenue transfers must account for who paid the taxes to build the current network.
As explained above, new mobility and autonomous vehicles have the potential to contribute to more dispersed land use patterns. Greater land use controls such as those listed below may be necessary to offset undesirable expansion of land use development.
- Urban growth boundaries – AVs and new mobility may extend the distance people are willing to travel between their home and major destinations such as employment and education centers. Urban growth boundaries are one mechanism for directing growth to help minimize undesired expansions of urban area footprints.
- Zoning changes – AVs and new mobility may increase development pressure on land areas and parcels that previously were not envisioned for residential development. Some cities and counties allow residential development under a wide variety of zoning classifications. With housing supply constraints in many major U.S. cities, AVs may extend travel distances as noted above, which could increase demand to build residential homes on parcels originally intended for other uses.
In sum, disruptive trends and new mobility have the potential to increase the use of vehicles and extend regional accessibility by lowering the costs (money and time) of vehicle travel. Offsetting potential undesirable effects requires government actions. It requires difficult policy decisions that conflict with the current norm. As we alluded at the start, and is reinforced throughout this article, the objectives of the private market to incentivize vehicle use may generate outcomes that are mis-aligned with important government objectives.
The private market will be incentivized to generate revenue from new mobility services based on miles of travel, minutes of travel, and choice of vehicle/service. This structure does not guarantee that these vehicles will be shared. Conversely it will likely increase trips, creating competition between service providers to reduce costs, increase choices and attract more riders. More miles and minutes mean higher revenue, and in this light, policy is needed to balance these private market interests with public policy goals. We provide some policy and regulatory responses for the public sector, but it is up to planners, engineers, citizens, and elected officials to take action. And that action should happen now. Onward.
See more in the forthcoming book “Disruptive Transport: Driverless Cars, Transport Innovation and the Sustainable City of Tomorrow” published by Routledge.
Leave your comment below, or reply to others.
Read more from the Meeting of the Minds Blog
Spotlighting innovations in urban sustainability and connected technology
A recent study by the International Downtown Association reports that vibrant downtowns contain around 3% of citywide land, but contain 14% of all citywide retail and food and beverage businesses, and 35% of all hotel rooms. This results in $53 million in sales tax per square mile, compared to the citywide average of $5 million. Not to mention that downtown residential buildings also add to the tax base. In the 24 cities included in the study, residential growth in these downtowns outpaced the rest of the city by 400% between 2010 and 2016.
Partnerships between city officials and contractors result in new and visionary downtown destinations. Along with large vertical construction projects, there are opportunities for countless other projects, including parking structures, enhanced Wi-Fi, landscaping, pedestrian and biking paths, and traffic improvements.
Ordered city geometry that is built today is meaningless for energy cycles. Resilient networks contain inherent diversity and redundancy, with optimal cooperation among their subsystems, yet they avoid optimization (maximum efficiency) for any single process. They require continuous input of energy in order to function, with energy cycles running simultaneously on many different scales.
Short-term urban fixes only wish to perpetuate the extractive model of cities, not to correct its underlying long-term fragility!
TDM, when employed, works. TDM agencies around the country use a treasure’s trove of strategies to get people out of cars and onto trains, buses, and bikes, which is something that has to happen if we don’t want our roads to become unusable due to traffic and environmental congestion.
But one major problem with the practice of TDM is that it has had a hard time making the case that it is a cost-effective alternative or at least add-on to big infrastructure projects. It seems pretty obvious that teaching people, educating them, about how to use our systems will make those systems run more smoothly. But there has never been a great way to back up that assumption with hard numbers.