Guest Post: Autonomous Vehicles and the VMT Problem

 This guest post is by Ron Kilcoyne, general manager of Lane Transit District in Eugene, Oregon and formerly the head of transit in Bridgeport, Connecticut and Santa Clarita, California. 


The flurry of speculation about the future of autonomous vehicles is mostly ignoring a signficant downside: the impact on vehicle miles travelled (VMT).  Safety and congestion resonate with people while VMT doesn’t.  Yet reducing per capita VMT is also essential for combating climate change. The potential increase in VMT when self-driving cars become prevalent could negate any congestion reduction benefit. Indeed it could be far worse than today.

Reducing VMT or per capita VMT is usually viewed as restricting individual freedom. But is it? If high quality pedestrian and bicycle infrastructure and high quality transit service is provided, individuals have attractive choices to driving and many use them. Add to this mix car sharing, bike sharing and transportation network companies (TNC) like Uber and Lyft and many households are choosing to go car free or reduce the numbers autos owned without sacrificing mobility and accessibility.

Much has been written about autonomous cars in this blog and elsewhere. Countless predictions on the impact autonomous vehicles have been voiced – and they are probably all wrong. One good analysis on the introduction of autonomous vehicles is here, Zipcar founder Robin Chase has a heaven or hell scenario here.

Autonomous vehicles will not eliminate the need for high capacity transit (and Jarrett makes this case here) and we should be concerned that opponents of transit investment will use the prospect of autonomous vehicles as a reason not to invest in transit.

But the real concern is on the impact on society and the communities in which we live. Will the trend toward walkable communities and more active transportation be thwarted? Will we become a more isolated society moving about in our autonomous pods? Will sprawl spread? And to the main point, how much will VMT growth inhibit efforts to combat climate change.  Even electric vehicles contribute to climate change if the electricity comes from fossil fuels, and fast growing VMT implies accelerated road construction, with its own environmental impacts.

No one knows how fast autonomous vehicle availability and use will occur. But let's focus on two of the most discussed impacts of the growth in autonomous vehicle penetration in urban areas:

  • reducing the amount of land devoted exclusively to the movement and storage of automobiles (a benefit) and
  • the increase of Vehicle Miles Travelled (VMT), causing inevitable pressure to pave more of the earth (a negative);

The former will only occur, and the latter will only be tamed, if we price the movement and storage of vehicles correctly.

Improper pricing of automobile use and storage has put public transportation at a disadvantage since the end of WWII and maybe even longer than that. There is plenty of literature that makes the case for proper pricing of road space, but as impressive as these arguments are there is little public support to increase the cost of driving or storing a car. Without public support there will not be political support.

Car sharing and bike sharing have been growing over the past few years. The prevalence of these options result in some households shedding an car and using alternatives, including using transit more frequently.  Indeed the two are mutually supportive. It is likely that households will reduce the number of cars they own (some may become car free while others will reduce the number of vehicles owned) if they have choices including a good car share program and good transit service.

If car ownership declines, and trips using shared cars are accurately priced, then individuals are more inclined to walk, bike or use transit when those modes are attractive alternatives.  High-ridership transit, in particular will still be cheaper than shared cars.  Therefore it is in the transit industry interest to partner, promote and facilitate growth of the sharing economy, but also to ensure that these services are priced fairly.  If so, the sharing economy and high-ridership transit should be the best of friends, offering complementary services at different price points and often connecting with one another.

Shared autos, whether in the Zipcar or Uber model are still not paying anywhere near the cost of using road space.  Once autonomous vehicles enter the picture, ZipCar and Uber will become indistinguishable.  Articles such as the one referenced at the beginning of this blog make an appealing prognosis about the potential to repurpose significant chunks of urban land but fail to acknowledge the impact of increased VMT.

If VMT grows dramatically we may find that the promises of freeing our cities of parking craters an illusion.   And for the same reason widening roads doesn’t end congestion – induced demand — the promises that autonomous vehicles will end congestion could be just as fleeting. Therefore the need to properly price road usage and parking spaces; and to include externalities such as greenhouse gas emissions (tailpipe for liquid fuels or at the source for electricity) becomes more imperative.

Any combination of greenhouse gas tax, vehicle mile traveled tax, weight –distance tax, congestion pricing, or variable tolls can accomplished this if priced properly. However getting any of these fees enacted is daunting. Opponents of ending auto subsidies are much more effective in framing the issue in ways that appeal to the average citizen. The human inclination to resist change and oppose paying are also on their side. Yet we have seen sea changes in societal attitudes happen in our life time – some take decades (smoking); others less than a decade (gay marriage). In this case we don’t have the luxury of waiting decades.  

Individuals and organizations that care about climate change, the quality of our urban spaces, and protecting open space need to brainstorm on how to frame this issue to build the needed political support to accurately price road and parking space usage. We can start by using the comment section by focusing on how we can accomplish this rather than giving reasons why we can’t.

To keep our cities moving and not overrun with vehicles, we need high quality transit in an autonomous vehicle world. Most modeling relating to the impact of autonomous vehicles indicates an increase in VMT even in scenarios that assume the availability of high quality high capacity transit. Unless the number of people arriving at a destination equals the number of people leaving at the same time there will be a lot empty autonomous vehicles moving about going to their next trip. If we can call up a vehicle wherever we are and, and if use of street space is heavily subsidized enabling these services to be inexpensive, induced demand will overwhelm our road network even with these vehicles travelling closer together.  

With parking taking up to 25% of the space in our communities large and small and road space another 25% or more, the prospect of reducing this is very appealing. One futurist claims we can reduce the need for parking by 80%. That may seem fantastical but even a 10% reduction in land devoted to parking can result in more green space, housing and employment opportunities in our cities without destroying the character of existing neighborhoods and allowing cities and small communities to grow preventing sprawl.

This can only happen if road space is properly priced, and if we invest in high quality pedestrian and bicycle infrastructure and high quality transit service. The question we need to think about and answer is how do we succeed in getting pricing and investment decisions right.

14 Responses to Guest Post: Autonomous Vehicles and the VMT Problem

  1. Paul Godsmark November 13, 2015 at 9:14 pm #

    If a municipality chooses to promote ride-sharing of fully autonomous vehicles (AVs), then this could have a significant impact on the number of vehicles on the road at peak periods, reduce congestion, reduce emissions (especially if the trend of AV development continues to focus on EVs). My own calculations show that if we can increase average vehicle occupancy during peak periods from the current 1.1-1.2, to around 1.8 or more, then the number of vehicles in peak periods can drop dramatically. Shared AVs then become a hybrid of transit and the private car, offering on-demand point to point service. The ride-sharing aspect pushing the cost of the service way below using Uber or conventional Taxi services and potentially on a par with conventional bus transit. The higher service/convenience levels of point-to-point service could see significant modal shift even if prices are slightly higher.
    We mustn’t forget also that we should be talking about autonomous VEHICLES that can come in completely new modes. Especially important to this discussion is individual pods that can collect passengers in the suburbs, city outskirts and then physically combine into road-trains to enter the urban/downtown cores. Thus the service levels of a taxi, but close to the passenger density and footprint of a conventional bus.
    An increasing number of studies support the reduction in vehicles required in ride-sharing AV fleets that are not noted in this article. In particular studies by MIT-Singapore, OECD (ITF) and the Earth Institute, Columbia University.
    We perhaps should not be looking at a VMT problem, but perhaps the increase in Passenger Miles Travelled as if AV fleets prove as successful as the above studies suggest then people will be making a great deal more journeys. In addition, the ‘forgotten 30%’ that don’t have a driving licence or can’t drive (many that are disabled, seniors, too young, too poor, medically at risk etc.) will now have much easier access to road transport. This group will generate a significant amount of new journeys that they otherwise find difficult or not possible in the current paradigm.

  2. Harry Audus November 14, 2015 at 5:31 pm #

    Increasing occupancy may well reduce VKT – I say “may” because any increase in effective capacity, on any congested mode, can induce more trips – but I don’t quite see how the advent of autonomous vehicles would encourage this. Ride-sharing is perfectly possible in current circumstances, and we already have taxis. Presumably the same limitations on use of autonomous vehicles would apply as apply now; that is, people may still like to own their vehicles and maximise their opportunities to travel whenever they want to, and autonomous taxis will still need to operate within a business model. Or will the absence of a driver (generally 50% of the fare here in Australia, excluding Uber) will make autonomous taxis cheaper to hire than current taxis? Furthermore, significant VKT are consumed by taxis dead-running, even if they don’t need to roam around looking for fares any more.

  3. CD November 14, 2015 at 7:54 pm #

    I look forward to jaywalking everywhere in the bold future where all cars will stop for me.

  4. David Marcus November 14, 2015 at 10:33 pm #

    There are two big benefits of autonomous vehicles which are glossed over in this piece:
    (1) Because parking lots are unnecessary, we are likely to see much denser development, which in term will actually be better suited to walking and transit.
    (2) Because the expectation is that people will give up their cars, the marginal cost of driving should go up, disincentivizing unnecessary driving.
    Both should have a beneficial effect on VMT per capita.

  5. P November 15, 2015 at 1:12 am #

    I didn’t really agree with this post. Lots of fear and uncertainty – the reality is that we cannot predict the future. If we look in the past, big changes don’t always have to be big disasters, and mostly aren’t.
    Road capacity will increase with the same amount of vehicles if the autonomous vehicles are shared. This would start to blur a line between the private car and the public bus. For example, a car lane with a car every 3 seconds gives a capacity of 1200 persons/hour (assuming 1 pax per car). Now if we can get 5 people in the car, that line capacity increases to 6000 persons/hour, an increase of 5x. That’s a HUGE gain in efficiency.
    In addition to this, the article confuses propulsion with mode. It’s a variant of the ‘electric trains are green because they run on electricity, buses are dirty because they burn fuel’ but applied to autonomous cars. It is possible to get diesel trains/trams and electric buses.
    There is no reason why cars cannot be clean, either through movement to electricity (and the electric source moving to greener methods) or by paying carbon offsets which remove the same amount of carbon that is output from the vehicle.
    Clean, autonomous vehicles will come, just like the horse and cart were replaced and a huge transition in the 1920s. Yes, the future has problems, but we have ALWAYS found a solution to them.

  6. P November 15, 2015 at 1:21 am #

    Oh, and a word about ‘sprawl’.
    What is ‘sprawl’ anyway? Is it really that bad? I tend to think of sprawl as the expansion of the city boundaries from the CBD. (Note: this doesn’t work for all cities, as not all are monocentric).
    There is a long list of evils attributed to sprawl, but I wonder if it is really about how far the city urban boundary moves from the CBD or the density of the development. Is it the actual distance that is bad, or is it the low density that is bad?
    Have these two separate attributes been separated, or are they just lumped in a big box togther that is then put on sinners’ parade.
    I think if the city boundary moves, but is sufficiently dense, that might not be a problem. You could also have more than one CBD (Sydney, for example, is polycentric with a recognised second CBD at Paramatta).

  7. Andre Lot November 15, 2015 at 4:45 am #

    I find this argument a bit confusing, but it goes in tandem with a stream of thought in the planning community that view autonomous cars (regardless of propulsion) as some Trojan Horse that could detract from larger goals (such as investment in transit and whatnot).
    Some 5 years ago, the suspicion revolved around the technical and financial viability of self-driven cars. Now that these seem to be a matter of ‘when’ instead of ‘if’, the arguments have coalesced around the evil autonomous vehicles can be as they would facilitate behaviors and choices dismissed as negative.
    One of the most naive assumptions, I think, is that people would just buy autonomous vehicles by the millions without a substantial shift towards non-ownership access. If you can have a combination of Uber-like (on-call, reliable and mostly prompt availability at the address you need) service that can be operated without drivers, costs would go down and the incentive for people to buy their own cars and let them idling 95% of the time would be lessened.
    I agree that autonomous vehicles are unlikely to be good substitutes for mass transit on major corridors, but they can be vital to provide mobility where, in current models, transit suffers, such as erratic errands outside central areas, late-night travel, last-mile connectivity much faster than walking/cycling without the need for massive parking garages at stations, taking children (that parents deem to young to ride transit) to choice schools outside the neighborhood and back etc.
    Even in cities well served by transit (think Berlin, Tokyo and London, for instance), trips that don’t follow the usual axes are far slower by transit than car. Many people who live car-free end up facing a huge time penalty if they want to do the occasional shopping or park visit that is out of their way.

  8. Jeff Wegerson November 15, 2015 at 1:03 pm #

    A lot depends upon ownership. What percentage of a particular city’s AVs are private owned versus publicly owned. And then communal private ownership versus individual private ownership. These percentages will matter when usage of limited road spaces becomes contentious. Like now.
    Cities that over-concentrate ownership in private hands, especially individual private hands have a more difficult time at making optimum use inside limited spaces. That results in less than optimum lifestyles in those areas. Where communal use is adequate then more optimum designed urban spaces become possible.
    The classic example is where existing autonomous vehicles run now and almost always as communal transit. Elevators.
    As AVs become smaller and smaller the pressure will increase to build elevators and offices where one can be driven right to their desk on the 50th floor. Likely the AV will then open up to become the desk/table and chair obviating the need for the passenger to get out to work or attend a meeting.
    A critical loss will be exercise. That will create the need for AVs to have built in exercise capabilities as found in a high-end gym.
    Laugh all you want but your option will, like the Amish, be to join various atavistic communities or “cults” that cater to the level of technology in which you prefer to live.

  9. John Levin November 16, 2015 at 6:15 am #

    While I tend to agree that autonomous vehicles will simultaneously increase both core urban density AND suburban sprawl, I think this whole discussion is beside the point when it comes to the question of how to address climate change. If the issue is Carbon, then address the Carbon. Whether it’s a carbon tax, cap & trade, whatever. If we attempt to manipulate the transportation/land-use systems without internalizing the externalities, we are bound to fail.

  10. Jim November 16, 2015 at 9:15 am #

    The problem with trying to predict the future is assuming that “all else the same” autonomous vehicles will do this and that. But autonomous vehicles will not evolve in a test tube. Most critically, it will evolve alongside autonomous package delivery vehicles which will be much smaller and consume much less energy as well as package delivery drones. A huge number of trips are made to simply pick up a very small package whether food, groceries, personal care items, etc. At the same time, many people will not have their own car that picks them up and then wastes a trip returning home to park. Autonomous shared cars will be continuously used throughout the day. I don’t see much need for downtown parking garages which will be replaced by residential and commercial spaces further increasing downtown densities. Let us not forget the most important benefit and that is less accidents. Humans were never designed to operate motor vehicles and 50K deaths a year is proof not counting pedestrian deaths.

  11. Murphstahoe November 16, 2015 at 10:20 pm #

    “If high quality pedestrian and bicycle infrastructure and high quality transit service is provided, individuals have attractive choices to driving and many use them.”
    Autonomous vehicles are the ultimate in high quality bicycle infrastructure. If you aren’t going to get hit… bingo!

  12. K November 17, 2015 at 4:16 am #

    While I think the promise of autonomous vehicles is overstated, this article does raise some interesting points. I’d never connected the possibility of road pricing and autonomous vehicles.
    Even though there is a public resistance to road pricing, there is some appetite within governments to move to a distance based charge for road provision. One of the technical issues preventing this is how you charge per mile. But an autonomous car will doubtless log exactly how far you’ve driven. It would make a move to per-mile charging extremely easy. That will get VMT down and people onto transit in no time.
    I do not expect to find autonomous car fans selling this as a feature any time soon.

  13. transportation services December 22, 2015 at 4:04 am #

    Nice blog, very well written about Autonomous Vehicles and the VMT Problem. Thanks for sharing.

  14. Peter Stark June 13, 2016 at 3:05 am #

    You are doing good work and important work here i appreciate you dedication and hard work when it comes Autonomous Vehicles and the VMT and problems of it . Carry on and keep working and we will keep visiting and seeing your work and maybe exchanging knowledge.

Leave a Reply