Tuesday, February 15, 2011

If users prefer rail to bus (or vice versa), what should be done about it?

Jarrett, writing once again on the theme of bus-rail differences (in preparation for a new book he's working on), writes an article on sorting out the rail-bus differences, along with a followup. In the series, he takes a list of (perceived) differences between bus and rail that were posted at The Infrastructurist, and sorts them into three categories: 
  • Intrinsic differences: Real technical differences between bus (including trolleybus) and rail (including, it appears, things like monorails and rubber-tire metros);
  • Misidentified differences:  Differences which are related to propulsion rather than guideway type, or differences which arise from how specific lines are implemented in specific systems
  • Cultural feedback effects:  Difference which arise from a result of cultural perception, law, or custom.
I won't repeat the list or the analyses here; go read the articles for those details.  However, there is a forest that is being missed for the trees.  Jarrett, I'm sure, knows this, but it merits pointing out.  The discussion of "cultural feedback effects" vs real vs miscategorized differences is important and useful, but there's a bottom line being danced around. That is:

1) There seems to be a quantifiable and measurable preference for rail over bus, at least in much of the developed world.

2) This preference may or may not be rational, may or may not be alterable (to the extent that planners ought to try), and may be conflated with issues dependent on local circumstances or practice, or who knows what else. (True apples-to-apples comparisons are often hard to make).

Given all of that, what should we (transit advocates, planners, decision-makers, or anyone else who cares about this) do about it, whether in planning individual lines or entire systems?  Specifically:

  • When and where should planners and decision-makers include modal biases in their forecasts (i.e. "a light rail line will attract X passengers per day, a BRT line will attract Y") and how much weight should these things be given?
  • To what extent should transit agencies, et al, attempt to change these views?
There are plenty of schools of thought on the issue.  One one hand, there's a school of thought that says that if reputable data (and proper decision-making requires access to good data) demonstrates a modal bias, that should be reflected in any subsequent analysis, and not colored by political considerations.

There's another school of thought that holds that such biases ought not be considered, especially if they are likely to be irrational (i.e a survey states that 20% more riders will ride on rail than on an equivalent bus line, and the stated reason for that preference is that busses are slow and full of drunks), and that only good "hard" reasons for one choice or another should be considered, such as cost or technical parameters, and that "illegitimate" reasons for a preference ought to be given no weight.  (This view is especially common when the reasons involved can be demonstrably reduced to racism or other forms of prejudice, which good public policy ought to oppose).

Some go so far as to suggest that modal bias (unjustified by tangible differences) is a problem that transit agencies ought to try and solve.

But these questions are ultimately political ones--decision-makers need to take all relevant factors into account.  I tend to favor providing as much quality data as possible to decision-makers, and letting them handle the politics.  (Not that they will necessarily do a good job here, but that is their job...) And while many non-intrinsic differences may be irrelevant in the abstract, they are often quite relevant in the context of a particular agency or system--if rail solutions gets signal priority at grade crossings as a matter of law whereas bus does not, or if a particular agency is not willing to add a new type of rolling stock to a fleet as part of a system expansion, these are important factors to consider.  If there are changes for the better in the wider environment (cultural, regulatory) that can be made, then plan for those; but transit systems have to exist in a particular place.

And difference which are specific to that place can frequently be just as important as those differences which are universal.


  1. I defend my schema from a long-term perspective, because it's rational to think long-term when thinking about transit infrastructure. The category of "cultural feedback effects" contains attitudes that seem overwhelming now but that may change over time, while the intrinsic differences won't.

    I agree that if your sole focus is the shorter-term timescale in which the actual decision is made, your analysis holds.

    But I'm trying to maintain a separation between prescribing and describing. Perhaps I should do a post on that ...


  2. Thanks, Jarrett. Obviously, the difference between prescribing and describing is important, and a distinction which is observed far too often in the breach. (Economics comes to mind).

    One good way to sort out this thicket, of course, is a finer scale, such as the Spectrum of Authorities attempts to do. Sometimes law is easier to change than culture, sometimes the reverse, and both are easier to change than physics and geometry, not that some policymakers don't try. :)

  3. In Portland, where I live, we have a fairly good bus and rail infrastructure. Some friends of ours from Salem came up to join us for dinner with their 2 year old daughter. They parked their car near the street car stop so they could ride to the restaurant where we were meeting them just because their daughter liked riding the train. So whatever the reason, the rail bias starts early.

  4. I blame Thomas the Tank Engine. :)

  5. Rail bias has existed in every culture which has had rail and buses so far, for the entire history of rail. And enough of it is due to intrinsic factors (things which are derived, one way or another, from the fact that a train runs on rails) that it's reasonable to assume that it's going to continue. The propulsion bias towards electric propulsion is equally real (though smaller in magnitude), for what it's worth.

    I think it's not actually worthwhile or sane to try to change it.

    Take a place where the trains are associated culturally with being slow and for the lower classes, and the buses are associated with being fast and for the upper classes. Like Turkey. This association can be reversed practically *instantly* with one good rail line, and it was.

    Take a place where it's the other way around: where buses are perceived to be worse than trains. You can only reverse this impression by making the buses very, very, very good, London-quality, while disinvesting in the trains, and that combination simply doesn't happen.

    Here's another thing Jarrett hasn't really thought about. Buses are more efficient (ecnomically) for low-volume routes, and scale up really really badly. Trains are not very efficient for low-volume routes, but scale up better than any other motorized form of transportation. This is not generally contested.

    But low-volume routes are inherently less attractive. The natural way to handle them is to make them low frequency -- which everyone knows makes them less attractive. Or, they can be high-frequency and low-volume, in which case their expense makes them the most obvious target for any budget cuts -- and if money is no object, as in a very rich community, then trains will be preferred for their ride quality.

    So buses can either be used in applications for which they are suited -- in which case they will immediately be seen to be inferior to trains -- or they can be used in high-volume applications for which they are not suited, in which case it will be even more obvious that they are inferior to trains.

    There's no way to avoid this except by deliberately running bad train services. When it comes to prestige, people won't compare the buses with the alternative of train service on the same corridor, they'll compare them to actually existing train services. And if you're doing anything remotely optimal, the train services will be in better corridors for mass transit and will therefore look better.

    Accordingly, I think the best bus advocates can do is to keep stressing that rail is for high volumes, that it just isn't worth the cost if you don't have high volumes, nice though it may be. Buses are cheaper when you just don't have the passenger counts for rail, and that's all there is to be said for them.

  6. neroden,

    I'm pretty sure Jarrett is aware of the capacity difference between bus and rail. And obviously, for low-ridership routes you want to send the smallest, cheapest vehicle possible; whereas for high-ridership routes you want the vehicles with high capacity.

    That said, there is an area of significant overlap--here, I'm focusing only on capacity.

    Take the MAX yellow line. It runs at 15 minute headways, both peak and off peak. Could the route be served by busses, if we assume that ridership would stay the same? Also assume that the bus in question has its own right-of-way. Certainly. During off-peak runs, the ridership of a Yellow train probably could fit on a bus (though it might be a bit crowded). During peak hours, frequency would have to decrease to provide a higher load, but one two-car train every 15 minutes is a capacity well within the reach of bus.

    What would the affects be?

    * Peak hour operating costs would go up, as one train would probably need replacing by 4-5 busses if you use 40' models (TriMet's standard), or 3 busses if you use articulated models.
    * Off-peak costs would go down, as a 40' bus costs TriMet about half as much to operate as a 2-car MAX train.

    What about customer perceptions of service?

    * Ride quality, arguably, would go down--fixed guideway transit has numerous advantages over free-running vehicles.
    * In the peaks, headways would improve. Many consider one bus every three minutes to be a better level of service than one train every 15.
    * There would be an option to have some of those busses branch.

    But, as you point out, ridership probably would go down, as there are a nontrivial number of riders who might use the Yellow as long as its a train but won't ride the bus.

    I'm not sure I agree with this:
    But low-volume routes are inherently less attractive. The natural way to handle them is to make them low frequency -- which everyone knows makes them less attractive. Or, they can be high-frequency and low-volume, in which case their expense makes them the most obvious target for any budget cuts -- and if money is no object, as in a very rich community, then trains will be preferred for their ride quality.
    There are many examples of high-frequency bus services that are not "low-volume". You don't define either term precisely, but I'm assuming high-frequency means 15 minute headways or better (being generous with "high frequency"), and that low volume means "bus nearly empty most of the time". Portland's frequent service network has many, many high-frequency bus routes which are well-patronized. And even though Portland has (arguably) money to build rail, there's no thought whatsoever of converting every bus line to rail--just major corridors.


Keep it clean, please