To Kill Fewer People, Rely More on Transit

It may seem an obvious point, but transit is a remarkably safe form of travel, especially compared to the private car.  A new report from the American Public Transportation Association (APTA) authored by Todd Litman puts some numbers to how higher transit dependence yields a transportation system that kills fewer people.

Public transportation is one of the safest ways to travel. It is ten times safer per mile than traveling by car because it has less than a tenth the per-mile traffic casualty (injury or death) rate as automobile travel. Public transit-oriented communities are five times safer because they have about a fifth the per capita traffic casualty rate as automobile-oriented communities. In addition, crash rates tend to decline as public transit travel increases in a community. Contrary to popular perceptions, public transit travel is significantly safer than automobile travel.

Credible research indicates that many planning practices that improve public transportation and encourage its use also tend to increase traffic safety. However, these benefits are often overlooked: individuals tend to exaggerate public transit risks; planners tend to overlook safety benefits when evaluating public transit improvements; and traffic experts seldom consider pro-transit policies as safety strategies.

It’s worth a peruse.  The press release is here, and the full report is here.

14 Responses to To Kill Fewer People, Rely More on Transit

  1. R A Fontes September 15, 2016 at 4:33 am #

    The most dangerous parts of a transit trip are getting to the first stop, time at a stop or on transit property but outside of a transit vehicle, and getting from the last stop to the destination. A shared autonomous vehicle (SAV) or robo-taxi eliminates these.

    • Matthew L. September 15, 2016 at 7:04 am #

      But all the strategies/technologies that would improve the safety of access, egress, and waiting (better sidewalks and street designs, better shelters with solar lighting, and CCTV and/or emergency phones at stops/transit centers) already exist. Shared autonomous vehicles don’t, and they would require just as much infrastructural rethink and political will to implement at a broad scale.

      Moreover, even though SAV’s would improve pedestrian and bike safety in the sense of being more aware of pedestrians and cyclists than the average driver, focusing on SAV’s over public transit would likely result in further deprioritization of safe sidewalks, bike lanes, cycle tracks, and appropriate neighborhood street design. SAV-using elites would consider them “not needed,” and SAV operating companies would have an interest in getting as many obstacles out of the way as possible in order to reduce risk and improve operating speeds. By contrast, focusing on public transit results in a greater emphasis on designing safe and appropriate street facilities.

    • Jarrett September 15, 2016 at 8:10 am #

      RA Forbes. A door-to-door service will never be cheap enough to be useful to everyone. Door-to-door is very inefficient for reasons are geometric, not technological, and are therefore not going to be changed by technology. I addressed this here:

      So perhaps my headline should be “Kill fewer people regardless of income.”

      • Alan September 15, 2016 at 9:49 am #

        Jarrett. With drivers, door-to-door will not be cheap enough – but whether that’s the case for SAV’s, which RA Forbes was discussing, depends on how much the vehicle costs to manufacture. Remember, it doesn’t have to be cheap enough to be useful to everyone – transit doesn’t meet that test today, which is one reason why we offer subsidized fares to large swatches of ridership and run coverage service. It just has to be cheaper than transit options. If that happens, it will be difficult to sustain transit outside of the areas that are truly geometrically constrained (like CBD’s and other very dense urban areas), which would result in a pretty severe retrenchment.

        • Kenny Easwaran September 15, 2016 at 5:28 pm #

          The cost of door-to-door technology in a dense area isn’t constrained so much by the price of the vehicle as by the price of the land the vehicle takes up. In Manhattan (or even the urban core of a city like Los Angeles) it will never be affordable for someone to get conveniently from one place to another by single-occupancy-vehicle.

    • Jarrett September 15, 2016 at 8:12 am #

      RA Forbes. Also, among the more devastating impacts of the “to your door” fantasy is the abolition of walking and thus the decline of interest in pedestrian infrastructure. This will kill even more low income people as they still have to walk to the anemic fixed route services that remain in this fantasy world.

      Thank God it’s only a fantasy, however expensive its PR!

      • R A Fontes September 15, 2016 at 9:58 am #

        I agree with the Columbia University Earth Institute report “Transforming Personal Mobility” that SAV’s should be cheaper than today’s cars because they would spread out fixed costs over more miles in a shorter period than typical personally owned cars. Also, the report’s suggestion that something like the Google “koala” should be even cheaper where appropriate makes sense.

        The biggest cost savings from SAV’s should come from the built-in capability to pool rides. Providers would have on their computers data on where riders are, where and when they want to go, and even some vetting from rider histories. Okay, it would be easy to get hung up on the traveling salesman problem, but fuzzy algorithms could go a long way towards quickly finding routings that are more than good enough, if not perfect.

        Unsubsidized SAV fares of less than 10 cents per passenger mile off peak really are possible.

        Yes, a full bus or rail vehicle is a better user of space than a bunch of full sedans & SUV’s. But we’re in a world where average carloads are less than two persons. If SAV’s increase those averages to two or more, won’t the world be a better place? Remember, too, that SAV’s won’t need parking space near destinations, but would just go on to their next missions or park well away from high density areas.

        Door-to-door doesn’t necessarily have to be dystopian. The primary market for SAV’s is not transit but the inefficient personal vehicle use. Researchers from multiple institutions have come to the conclusion that an SAV can replace about 10 personally owned cars. The reduction and possibly even abolition of parking minimums is a logical outcome of at least some car owners switching to SAV’s. Take out the driveways & garages, make the streets narrower, and residential lots can be smaller, cheaper, and yet offer more usable space. Suburban town centers could transform themselves from being dominated by oceans of asphalt to places that welcome walkers and cyclists while enabling community activities unthinkable in a world of massive private parking lots. It all depends on public policy.

        • Marc September 15, 2016 at 4:33 pm #

          But doesn’t it seem kind of pointless comparing an existing, mature transportation technology with a hypothetical one that hasn’t even scaled up for mass use yet?

          There are only a literal handful of test autonomous vehicles out there. Once/if they actually scale up to serve a large enough base of people to begin compiling statistics, then we can start comparing them with transit.

          In the meantime we might as well compare transit with teleportation, which is an even safer way to travel!

          Also don’t overlook Jevon’s Paradox when hypothesizing over the efficiency of SAVs (for example, requiring less road space). There has never yet been a transportation technology other than space-efficient public transit that has reduced urban congestion: easier/cheaper travel always ends up eating away any initial space and capacity savings. For example, I can indeed see SAVs reducing the number of single-occupancy vehicles someday, but if empty SAVs in urban areas end up cruising between tasks or going off to idle in parking lots “on the edge,” then there goes that recently-freed-up road capacity…

          • Alan September 16, 2016 at 9:10 am #

            I suspect that if teleporter technology was at the point where not only did teleporters *exist*, but they had progressed to the point where they were being commercially introduced by private companies to live customers for field testing, we would be talking about little else.

            My metro has a development and infrastructure planning horizon out to 2030. By that time, it’s likely that autonomous vehicles will have caused non-trivial changes in rubber-tire transit. Probably in at least the suburban areas, off-peak users, and through the initial adoption of driverless buses; but perhaps even starting to pick off some suburb-to-core commuters. That may not seem like much….but in a metro where the governmental bodies that run transit are elected from *districts*, de-emphasizing transit in the suburbs and losing all those transit workers will seriously change the political calculus behind funding fixed-route transit.

          • Marc September 18, 2016 at 10:14 am #

            Hm, I’m not sure that simply having a handful of carefully-scripted, Potemkin-style AV tours guarantees that the technology can scale up for mass use, particularly in the optimistic timeline many are touting.

            I’ve no doubt we will eventually overcome the many obstacles to true autonomous driving…
            …but the problem remains that it seems we’re basically asking computers to function as social beings.

            The heavily-scripted and carefully-programmed AV tours we’ve seen so far are a far cry from true social spontaneity: they’re more like intensely-planned and carefully-mapped presidential motorcades.

            Just because a handful of science projects are being tested is no guarantee that mass use is around the corner. I remember looking through reams of old Popular Mechanics and laughing about all the promising new technologies in active testing that seemed just around the corner but never actually scaled up. We seem to suffer from a technological hubris in thinking that every new thing is guaranteed to go big, but for every technological breakthrough that has indeed gone mainstream – drones, for example – there are scores of others – like EVs – that never seemed to go mainstream despite significant testing, development, and even limited public adoption for centuries.

            AVs are exciting, but I’m happy to wait and let them go mainstream first before speculating how existing transportation technologies should be remade in the image of a technology that has yet to scale up.

            I guess I’m one of those “late adopters” – for example, I didn’t get a smart phone until the technology was cheap(er) and mature. I preferred to let that giddy first wave suffer through the initial shortcomings and limitations! 😀

          • Alan September 19, 2016 at 9:05 am #

            Perhaps it won’t scale up. It may end up being too expensive in the near term to make AV’s work for personal transportation – after all, there have been devices (looking at you, Segway) that never really made a market for mass adoption. And there’s nothing wrong with *personally* waiting until new tech goes mainstream.

            However, if you’re in the transit planning business, you might not have that luxury. Along with freight, for-hire transport is looking like one of the “bleeding edge” segments in which AV’s first get adopted. And your planning horizon is more than a decade out – you have to make decisions *today* based on what you think 2030 will look like. So if you just wait and see, you may find yourself making some poor resource allocation decisions over the near term.

            We don’t know yet whether for-hire AV transport will be cost-competitive with transit *generally* – but Jarrett’s “geometry” argument suggests that economics aren’t all that relevant, because AV’s will be blocked from substituting for transit by space limitations. I don’t see that as very compelling outside of a handful of metros with very high transit usage (the major NE metros, Chicago, SF, and maybe Seattle) and perhaps the specific CBD’s of a few others. In most other metros (and especially for non-CBD and off-peak service) the mode share of rubber-tire transit is so low that switching their riders to cars will have a negligible “geometric” effect on road usage.

  2. Erick September 16, 2016 at 7:52 am #

    Regarding the door to door dystopia. I am scared also of the consequences to our roads and therefore on public treasuries of all those vehicles not needing parking in a CBD.

    Sure they might return to the outskirts to pick more passengers but the reality is that they will drive back to the outskirts to park and wait to deadhead back to the CBD in the next rush hour. More vehicles more traffic more expenses on road upkeep.

    That along all the other points raised by Jarrett and other commenters. This fantasy needs to be stopped!

  3. Bjorn Swenson September 16, 2016 at 11:12 am #

    Better Friday and Saturday night service likely produces the greatest marginal reduction in traffic deaths.

    Labor: In my experience, transit operators tend to view weekend late night shifts are less desirable and usually leave them for the drivers with the lowest seniority. Bus companies tend to have to pay shift differentials for late night shifts, either as a requirement by the union or out of necessity. My guess is that an operator is scheduled to always work a late night shift, they are more likely to call in “sick”. At my agency, drivers are randomly assigned one weekend late night shift per semester; this is a fair compromise that distributes admittedly more challenging work equally while keeping somebody from having to give up every weekend night.

    At the same time, there are also people who would prefer part-time late-night work. People with a 9-5 job but are willing to moonlight on weekend nights for extra pay are potential workers, college students like myself are another. (My opinion is that city transit agencies should seek out college students for part-time bus operator positions, but that’s a whole other story.)

    Effective late night services require a different service approach than daytime routes. Like it or not, passengers are especially unwilling to stand on a corner of busy arterials while waiting for their connecting bus to show up or dart across an arterial only to have to walk past a closed, poorly lit commercial area; women are especially likely to have this concern. Thankfully, late night demand is ‘one-to-many;’ from the entertainment venue to home. In smaller cities, late night origin demand is concentrated in a handful of places, usually the downtown and near the universities.

    A timed-transfer, hub-and-spoke service pattern maximizes passenger comfort and safety. Smaller cities can get away with a hub downtown or near the university, larger cities will need suburban hubs. These hubs should be off-street, well-lit, and routinely patrolled to maximize safety; routes should also be scheduled to keep transfer times short.

    Late night passengers are less time sensitive and more safety sensitive, making curbside service an excellent choice. My agency uses 30′ Gilligs for some paratransit shifts; in my experience, 30′ buses are very nimble and are easily fit down suburban residential streets or through apartment and business parking lots (as a rule of thumb, if a garbage truck can fit, so can a 30′ bus). 30′ buses provide plenty of capacity for suburban routes while also being nimble enough for the agency to offer curbside service. As late night service would be very effective in reducing traffic deaths, acceptable route productivity (to me) is lower; while daytime suburban routes are generally unproductive below 20 passengers per hour, I personally would find service with perhaps only 10 passengers per hour to be worth the mortality gains.

    • Ian Mitchell October 5, 2016 at 10:23 am #

      I agree with you that keeping the drunk people off the road is the lowest-cost way to reduce road fatalities.

      I think that it’s one more easily served dollar-for-dollar by barring bars from providing parking, and levying an alcohol tax which matches uber/lyft/taxi (ideally uberPOOL and other shared services) fares- so a $20 ride is $10 to the drunk (or “late night”) rider.

      This (the latter not the former) was implemented in my hometown. It works.
      Plus, a totally distinct nighttime system is confusing to regular riders and irregular riders both, especially if they aren’t sober.