No, Let’s Not “Uber” Our Bus System

Have you just read another article claiming that public transit would be better off if we unleashed private innovation?  Ask whether they're talking about privatized operations or privatized planning.  These are totally different things, but it's currently fashionable to confuse them.

Graduate student August Ruhnka, writing in the Denver Post, is the latest in a series.  After reviewing the real cost and quality control issues plaguing US bus systems, he goes on to propose a fatal confusion between privatization of operations and privatization of planning.

The remedy is simple. The city awards a route bus contract after a competitive bidding process, thus entering into a private/public partnership.

So far so good.  This is the standard model of privatized operation.  But then, OMG:

By privatizing our bus system's day-to-day operations while still providing public oversight, private operators will be able to change routes, schedules and fares as frequently as necessary without the need for lengthy public hearings and political approval.

"Change routes, schedules, and fares?"  This is the privatization of planning.  It features the abolition of "political approval," which is a polite term for the abolition of democratic controls over spending the public purse.   It's also produced some of the worst bus systems — in terms of both disutility and waste — that I've ever encountered in 23 years doing network planning.  Ruhnka goes on:

Today, successful privately operated transit projects like the Sydney Metropolitan Bus System, Hudson-Bergen Line in New Jersey and the JFK Air Train in New York are chugging away.

Only one of these examples is even a bus service, where route changes are an issue, and none of them are examples of the privatization of planning that Rohnka proposes.   They are all examples of privatized operations.  

The Sydney example is more apt than Rohnka knows.  Planning was controlled by private operating companies over much of Australia and New Zealand when I started working there in 2006.  Now, all over both countries, planning is being moved back into the public sector, while operations remain private. That's happening, in part, because transit planning by operating companies produced fantastically inefficient bus systems such as the one in Auckland, New Zealand, which I have been helping the new Auckland Transport to redesign.  

The problem was not just privatization but turf.  Private companies who control just certain bus routes may optimize their own bus routes, but that doesn't optimize the whole taxpayer funded network and can in fact make it worse.  (The same problem can afflict turf-bound governments, of course, but turning them over to the private sector does nothing to solve that problem, while it does remove our ability to hold elected officials accountable for it.)

In cities across the world, major bus corridors are becoming important redevelopment areas.  Bus service has a huge nexus with a range of important government activities, notably road design and land use planning.  To give up control over the design of bus services is to give up control that's needed to do those jobs well.

Public hearings are a drag, and the approval process for bus service changes in America is exhausting for both elected officials and the public.  Preparing for this process is central to every project I do.  My job is to help elected officials (or their appointees) make clear decisions about the real tradeoffs that transit planning requires.  These are hard choices that must reflect a city's ambitions and values, and that must be linked to a city's decisions about land use, urban form, and social policy.

All these thoughts are supposed to give way, of course, to the romantic idea that we could "Uber" our bus system.  

In light of our Uber-dominated downtown, it is time for us to make this switch. Private-route contracts establish a sustainable procedure to constantly test the market to achieve the lowest cost. Commuting in downtown Denver requires three things that currently RTD lacks: demand-response (paratransit) operation; experimental services; and remedying routes that have low ridership. All the above situations require a level of flexibility that the private sector can provide.

Uber is big and new and financed and sexy, so how could this not be better than our old bus services?   (A dismissive view of  "political approval," too, is part of the Uber mystique.)

Uber today is a taxi and (limited) shared ride service using small vehicles to carry small numbers of people at once.   Assuming you pay the driver decently (a big assumption in the private sector) the cost-effectiveness of transit is going to lie in passengers per driver, because pending driverless vehicles, the driver is most of the cost.  And the only way to get that number high is to run large vehicles in fixed-route services that are so well designed, and optimized over so many purposes, that lots of people ride them.   The Uber model does not scale to large-vehicle fixed-route transit, which is also known as cost-effective transit.  

Why, then, does downtown Denver require "demand-responsive operation"?  Downtowns are big, dense, and need to use street space efficiently, so large vehicles are the key.  Compared to big buses with decent ridership, demand-responsive service is a way of carrying very few people at a high cost.    There is no way that a demand-responsive solution, in a place where fixed routes could work just as well, makes any sense as a way to make transit affordable to low-income people, one of Ruhnka's alleged concerns.  Affordability is scalability.  If a solution doesn't scale efficiently — in terms of labor cost, energy, and urban space, it will naturally be expensive.  

So to the extent that some people think they need a demand-responsive service downtown (apart from paratransit for the disabled) by all means let Uber and Lyft do that.  Demand-responsive is such an intrinsically inefficient form of transit that deploying it downtown can only be for the purpose of serving relatively fortunate people at fares much higher than transit fares.  That's a great role for the private sector.  There is also a role for demand responsive service in suburban areas where development patterns preclude efficient transit, through contracts between demand responsive providers and transit agencies.  But not downtown.

Another way of describing all of the "demand responsive" or "Uberization" fantasies is that they are predicated on moving large amounts of steel and rubber per customer trip, compared to big-vehicle fixed-route transit.  Even without considering the economics of labor, this can only be less efficient, in terms of energy and urban space, then what crowded big-vehicle transit achieves.  (And yes, you've probably read that a lot of big buses run around empty, as though this means that buses are a poor tool.  The refutation of that argument is here, and here.)

Transit is full of opportunities for private sector involvement, including in operations, infrastructure, and my own job, professional advice.  The private sector is always welcome to innovate, and there are markets — generally for higher-end services and higher fares — that the private sector will probably take over.  There are also fascinating opportunities to use Uber/Lyft models to improve the efficiency of services to low-demand areas — that will require partnership with suburban communities or transit agencies.

But privatize planning?  Let private companies re-arrange transit services without regard to the impact on the city and its values?  That's the opposite of democracy.

the evolution of logic in privately planned transit

Step out into most developing world cities, and you'll see something like this:


 Lots of vans sitting around, or looking maybe like they're about to go somewhere useful.  Vague cardboard signs in the windows suggest they may or may not be public transit of some kind.

They're called mutatus in Nairobi, colectivos in Latin America.  Over much of the world these informal, private, for-profit vans, run at low cost for low fares in areas of high demand, forming the basic public transit for a city.  Generally they run along a particular route out of a hub like the one above, but sometimes it's possible to vary the route depending on what you can negotiate with the driver.  You can count on them to hit key locations but not necessarily the exact path they'll take.  You also can't be sure of when they'll go.  Sometimes they wait until they're full before leaving.  

Today in Atlantic Cities Emily Badger tells the story of the Digital Matatus project, an attempt to map and describe the spontaneously evolved patterns that these semi-fixed-route buses operate.  Although nobody planned this network, it's more orderly than you'd guess.  Download the sharp, complete map here.

Matatu map slice

What do I notice?  Practically everything goes downtown!

Matatus have organized themselves into routes because that's to their benefit; they train customers where to wait for them along reasonable paths so that they aren't driving around looking for customers individually.  The idea of the route — and of an efficient, non-duplicative spacing between routes — arises spontaneously from the economics of the product.

But they almost all converge on downtown, creating huge jams there.  Nairobi is clearly big enough to  have large flows of people crosstown to many non-downtown destinations, suggesting that a more efficient and liberating network would have more grid elements.  This is a common thing that goes wrong in privately evolved systems.  Every matatu wants to go downtown because it's the biggest market, and a mutatu driver doesn't have to be coordinated with anyone else to fill a bus going to and from there.  This  problem that bedevils privately routed and scheduled operations everywhere.  

Crosstown service, by contrast, requires frequency on a single path connecting several major dots, and it has to leave from organized non-downtown hubs where many other services connect to it.   That requires more organization, so it's less likely to arise spontaneously out of private operators optimizing for themselves. 

So you get a single market overserved and other markets underserved.  This is very much like the way a narrowly-focused transit agency will throw too much service at a single market rather than building a network useful for many markets.  It takes more planning and management to create a network, and this usually requires a government willing to impose order.  

This same problem was observable after the wholesale privatization of buses in Britain.  Suddenly there was lots of duplication of bus service into the biggest downtowns as everyone chased the easiest prize, but service disappeared from crosstown markets that could have done well, but that required a network of organized connections to succeed.  That network is what privately motivated transit has trouble delivering, because it usually requires cooperating with people who are perceived as competitors.

Now and then, these systems get reorganized by government into more logical routes that spread the network across the city for easier everywhere-everywhere travel, as happened in Santiago in 2007.   The transition is hell, but when you're finished, you have a network that's much easier to use to go all over the city, and a much smaller knot of buses downtown.

The moral?  Disorganized transit systems "planned" by the actions of many private actors do naturally evolve certain forms of efficiency, but they do not naturally evolve into the most efficient and productive network for the whole city.  That final push into coherence requires network design!

guest post: shaun cleaver on zambian public transit

Cleaver is a PhD student at the University of Toronto. His main career focus is
disability and rehabilitation in low resource settings. This work has taken
Shaun to Haiti, remote northern Canada, South Africa, Cameroon, and most
recently Zambia where he is exploring the possibility of conducting
participatory research with leaders in the disability community

I am a
temporary resident of Lusaka, having recently relocated to Zambia’s capital
city from Canada. Now that I am here I need to get around, and doing so has
been a voyage of discovery into the public transportation system that is
responsible for most of my comings and goings in this sprawling African city.

There is currently no system to
disseminate information on public transportation in Lusaka.  As a new rider
looking to understand the network in its entirety, I have been forced to cobble
together information gleaned from specific discussions, my own experience and
observations, and rare nuggets left on the internet like messages in bottles
left to float on the great Internet Sea and hopefully find their way to future
adventurers trying to make sense of the chaos.

analyzing operations here I am building upon my perspectives as a regular user
of multiple systems in high-income countries (particularly in Southern
Ontario), as well as those of low-income countries where I have lived, such as
Haiti. Where they apply I will draw upon the principles described by Jarrett in
this blog and the associated book. As Jarrett states in the Introduction of the book, there
are some important differences between public transportation between “the
developed world” and “the developing world,” but also some common phenomena. I
will use this post to identify the characteristics of such a system to provide
a baseline perspective for blog readers unfamiliar with these realities. Indeed, some analysts suggest that these features
should be applied to transportation systems in high-income
countries too
, making it even more relevant that riders in those countries
understand the consequences of such structures.

Loading: What makes it all go?

Like many
cities in low- and middle-income countries (especially
), public transportation in Lusaka is operated as a seemingly infinite
number of mostly-independent small businesses that depend exclusively on fare
revenue. The backbone of the system is the minibus: a van with row seating
operated by a 2-person crew (a driver and a conductor, to whom I will refer to
collectively as the operators). I have heard differing accounts as to the
proportion of buses that are operated by their owners – as compared to those
that owned by entrepreneurs in some revenue-sharing arrangement with the
operators – leading me to conclude that both models are common. In order to earn a living the drivers and conductors need to maximize the revenue from their one vehicle
while minimizing the operating expenses, the most substantial of which is fuel.
With massive unemployment in Zambia and low wages the norm for the masses,
there are many people willing to do this work on a rather tight margin. Time is
a concern for drivers, but less-so than the cost of petrol, so the constant
preoccupation of the operators is ensuring that all available space on the
moving vehicle is earning fare revenue. Usually the bus will generally not move
until full (the exception being when movement is likely to help it fill).

At this
point I feel obligated to substantiate what is meant by “full”. In objective
terms, this means that each of the benches in the four rows behind the driver
has four fare-paying adults, and that there are another two adults in the front
next to the driver. Thus, the minibus is only “full” when there are 19 people
in it (and it is possible to squeeze another few riders if there are low odds
of a police checkpoint). Remember that I am referring to a vehicle that is
effectively a van. Consider this: when boarding I usually prepare the fare
prior to entering, as when I take my seat I am not able to reach into my pocket
due the proximity of the other riders. 

addition to the “market forces” that incentivize the individual operators to
pursue ridership goals, the system has
regulations imposed upon it by the municipal and national governments, and some
amount of collective self-regulation. The governmental regulations apply
primarily to the vehicles, but also to the routes and stops. The most visible
aspect of those regulations is the mandatory colour scheme, which has
traditionally seen all registered public transportation vehicles painted
different shades of blue and white (although white with an orange stripe was
recently approved as an acceptable alternative). Vehicles are also registered
by the Road Traffic Safety Authority (RTSA) as having met certain safety

have varying accounts of the regulation of routes and stops, but the evidence
is pointing towards minibuses only being authorized to and pick up and carry
passengers on a limited number of pre-identified roads. On certain major arterial
roads the stops are more clearly established with designated pull-out areas,
whereas on other roads it is common practice to pick passengers up or drop them
off just about anywhere.

Photo 2.1

Fare chart displayed on window of minibus.

Operators have organized into a syndicate, but it is not clearly visible to a rider and seems to be only rarely referred to in the
media. Nonetheless, operators seem to have a collective voice to negotiate the
designation of roads as being minibus approved and to set fares. Minibus fares
are formally established and publicized according to a fare by distance model. It is notable that the publicized fares do not include every
origin/destination possibility, yet every journey has an established and
precise fare that is collectively known to conductors and regular riders,
although often rounded up or down to the nearest half kwacha.

Of note,
journeys that start and finish among more quiet parts of the routes seem to be
priced more affordably than trips of similar distances in busier areas: I am
unsure if this is the product of intentional calculation, or simply based upon
the collective experience of what riders will pay and operators will accept in
order to fill the minibus. On the routes where large buses run these sometimes
cost less than minibuses, although the ride is generally slower due to the
longer fill times.

Setting off: How it all fits together

urban form and road network converges with the imperative to maximize ridership
to create a radial system focused on the traditional commercial area west of
Cairo Road (referred to locally as “Town”), where the minibuses serve four
central terminal stations. According to these conditions the radial system
self-perpetuates. Riders know that any destination in the city can be accessed
through a connection in Town, and therefore usually head there unless there is
a specific outbound destination in-mind. Operators want their bus to be full
before moving and the sequential departures in the downtown terminals ensure
timely filling. Once a vehicle sets out on it route it will likely have
passengers heading all the way to the route’s outer terminus; there the
operators find that the best way to fill the minibus is to serve the most
popular destination – Town.

In a
system that is driven by millions of individual micro-decisions there are few
examples of system-level thinking, making information on the system as a whole
conspicuously absent. In trying to understand the collection of possible trips
that the system allows I have had to patch together multiple practical
questions about how to get from individual points A and B. Almost inevitably
the answer goes “First you look for a bus into Town; when you get to Town ask
for the bus to where you are going.” Lusaka's
minibus fare structure includes no provision for making connections between
different routes. Instead, the rider must pay a fare penalty with every
boarding. This has the effect of enforcing the "one transfer into
town" model as the cheapest way to get from point to point, even if making
connections would effect a time savings or shorter trip distance.

Photo 3.1

llennium Station, the smallest of the four central terminals, on a quiet Sunday afternoon.

To date I
have only seen one example of a route map; one rider’s best guess at
identifying the patterns of movement as drawn onto a Google Map. Consistent with my
experience, the legend is laden with question marks; fortunately, it is close
enough to being accurate as to be useful. Other points of reference that I have
found include an ode to the
minibus operators
 and a superficial but
practical account
that essentially concludes that ‘it’s really just far easier to
take a taxi.’ 

this seeming anarchy, there are clearly routes. The picture of Millennium
Station shown here demonstrates the (surprisingly) orderly system of organizing
the minibuses by destination, a welcoming particularity of the central
terminals. As a new rider it was initially nerve-wracking that the vehicles
themselves were void of visual markers to indicate the route. Having been a
Lusaka-resident for less than two months I have already internalized the
irrelevance of such markings: if the minibus wants you as a passenger they will
let you know where they are going. If I am headed home but not at one of the
downtown terminals I know to listen for a conductor hanging out the window
yelling the familiar “Garden,
Ng’ombe, Roma,
 yooooo!” that will take me along the predictable route

readers will have presumably identified certain drawbacks of this
organizational arrangement. “If all of the minibuses converge on a limited
area, does this not strain the road capacity?” (Answer: yes, yes it does.) “If
there is sufficient demand for origin/destination pairings outside of Town,
would drivers not seek to fill that void and serve those riders?” (Answer: sort
of, read on.) “If all of the minibuses fill at the terminal, what happens to
those riders who try to board minibuses along the route?” (Answer: they watch a
lot of full buses drive past them while waiting for one that stops to let off a
passenger; or, for reasons to be explained shortly, a spacious minibus could
pull up, but this is usually not as much of a blessing as it seems).

Photo 3.2With
almost all vehicles converging on the city centre the congestion there is
indeed horrendous. For most of the connection-required trips that I take I know
in advance that up to half of the travel time will be spent inching down one of
the few roads into Town, before waiting for my next minibus to fill and inching
back out again. For this reason it is very desirable to identify travel possibilities
that do not include a downtown connection. 

I have found a few. One of these is at the terminal in front of the University
Teaching Hospital (in the central south-eastern part of the city, well outside
of Town). Here there are minibuses that serve other parts of the city using
orbital lines of travel, including two routes that get me close enough to walk
home. Using either one of these reduces my travel time by a minimum of 30
minutes. Sounds great, eh? Sort of – if we bear in mind some caveats. The first
is that the minibus takes longer to fill; and only does so reliably at certain
times of day (particularly as the hospitals day activities come to an end,
around 5pm). The  time I save traveling
is sometimes more than accounted for while sitting on a slowly filling bus for
an hour. Next, the operators charge a fare premium for the “short-cut” (a term
that seems to nearly gain official status in the frequent fare disputes among
unsuspecting riders). To be fair to the operators, the fare premium is not
unwarranted due to the time lost in filling this lesser used route, and the
operators seem to have to pay a fee to use the terminal: serving these unusual
routes does indeed address an unmet demand, but does so at a cost. As a rider I
am still in the process of determining when it is in my interest to use
“short-cut routes” and plan my journeys accordingly.

the use of the terminals seems to be a calculated decision on the part of
operators. In one sense the benefit is clear in that the vehicle at the front
of the queue is guaranteed to fill and then the driver and conductor need only
to replace disembarking passengers along the route. On the other hand, time is
lost while waiting for one’s turn and the fee to use the terminal cuts into

For these
reasons it is not uncommon to see a minibus begin its journey somewhere along
the route. As a passenger, however, boarding an empty minibus en-route can
prove to be a tactical error that often leads to a frustrating wait of unknown
duration as the conductor runs up and down side streets looking for any signs
of potential paying customers. To give the minibus the appearance that it is
about to leave the drivers will often turn on the ignition and inch the minibus
forward as the conductor frantically tries to steer people on “Come, come,
let’s go! We’re going!” moments before the driver cuts the ignition once more.
Often enough the bus will venture off-route as the driver and conductor whistle
for attention.

occasionally the vehicle will proceed forward along the route to repeat this
ritual in more promising locations. It sometimes, but rarely, occurs that this
forward progress is sufficient to allow a rider to reach a destination before
the minibus fills.

Photo 3.3

Inbound minibuses on the Ng’ombe-Town route. The minibus in the foreground is following the line of the route. The vehicle in the background (with its door open) was previously on the main road, traveling left to right. The operators chose to stop at the side of the road to collect more passengers and upon seeing none began reversing up the side street (the vehicle eventually disappeared from view; I left before it emerged).

On the road: the pearls and pitfalls of this type
of system

In this
environment, where the city is large and private automobile ownership is beyond the means of most, there is substantial demand for public
transportation. The capital investment required to serve that demand (by
purchasing a minibus) is notable but not extravagant. High unemployment and low
wages mean that there are plenty of people willing to depend on the thin and
unpredictable profit margin earned by drivers and conductors. The inevitable
response to that equation is the abundance of
which makes competition a necessity – but not usually in ways
that make operations more pleasant.

feel one perpetual manifestation of this when approaching busy mid-route stops.
Since I frequently board inbound minibuses at the University of Zambia I happen
to know that stop to be notoriously undesirable: accessing the stop requires
walking along a long footpath with only one destination, so the conductors
start aggressively courting passengers over a hundred metres from the stop,
well before riders can see which vehicle they are being asked to board. Once a
conductor has claimed a passenger for his (usually empty) minibus there is an
understanding that the passenger belongs to him. Of course this conflicts with
the interests of all competing operators (and quite often with the interests of
the passenger) leaving physical aggression as one of the few tools to enforce
the claim. The conductors often shout at one another and try to steer
passengers towards their vehicles. Sometimes the situation comes to blows. The
entire charade occurs while there are more than sufficient passengers to fill
any one given vehicle, yet a multitude sit immobile as operators calculate how
many more passengers are worth their while to compete over. 

At busy
stops and some terminals there is another group of individuals who add to the
dynamic, the “call boys” or ngangwazi. Call boys provide the
“service” of collecting passengers for approaching vehicles, in exchange for a
payment from operators. Readers familiar with the informal economy in
low-income countries will know that this is not merely a benign value-added
service: the call boys will steer riders away from operators who do not pay.
Call boys can be very aggressive in their activities as their earnings are
based upon their ability to influence ridership and operator behaviour, and
aggression is one of the few tactics they have at their disposal to achieve
this goal. Call boys are thus generally disliked, although their presence is
usually tolerated as a fact of life. On occasion the activities of call boys
can reach a tipping point that stimulates an organized response, as has
recently occurred at one of the terminal

Photo 4.1

Above: Showgrounds-Manda Hill bus stop on Great East Road on a Saturday. This is pull-out stop design is typical on Lusaka’s arterial roads, although some have the additional feature of a barrier between the stop and the road with a single entrance on the approach side and an exit on the far side. Note that this photograph was taken from the sky walk over the intersection of Great East and Addis Ababa roads; a maddening piece of pedestrian infrastructure that (along with the associated fencing of the intersection) forces one to climb/descend two stories to cross the road. On the plus side there is at least a reliably safe passage, a feature not available at the bus stop.
Photo 4.2

The large-size bus that I was riding was full and ready to depart the limited access pull-through stop at Millennium Station. Just before it left the minibus pictured at right reversed in through the exit to block our passage and collect a few additional riders.

designated pull-outs that have been created for stops are at least advantageous
in that waiting passengers are more fully separated from the speeding traffic.
Those with specific entrances and exits (presumably designed to instill flow on
the otherwise erratic vehicle movement around stops) can occasionally be a
curse, however, as it is not uncommon for a minibus to bypass the entrance and
park blocking the exit – ensuring that no one moves until that (now) front
vehicle is full. Predictably, this strategy also
comes with conflict.

corollary of the empty minibus waiting to fill is the full minibus speeding
toward its destination. Many of the secondary roads in Lusaka have wide rights
of way but only one paved lane in each direction. When traffic is heavy it is
thus common to see drivers leave the road in order to bypass traffic. Although
I have yet to see a vehicle stopped for this manoeuvre I am quite confident
that it is illegal. Interestingly, this is a strategy that is mostly employed
by minibuses and the occasional taxi, making me believe that this is a
calculated risk on the part of drivers where the potential of a fine/bribe is
weighed against the revenue lost by weighting in traffic. Using the logic of
spontaneity, it is almost possible to interpret the unpaved shoulders as public
transportation queue jump lanes. 

Lusaka public transportation vehicles arrange riders by rows they are equipped
with “flip down seating” where the aisle disappears as the vehicle fills. This
design means that half of the bus must disembark/re-embark if a passenger in
the back right corner needs to alight, a reality that passengers seem to accept
readily, although grudgingly. Minibuses thus fill in a predictable pattern
where the permanent seats near the front fill first, followed by the back seat,
followed by the aisle. Although it is tempting to try and avoid the most buried
of seats it is necessary for someone to take them – and the vehicle will not
move until someone does. Besides, sitting in an aisle seat is no relaxing nap
either as the rider needs to regularly exit and re-enter the vehicle to allow
others off.

Fares are collected by the conductor in a wave starting at the
front and moving backwards as soon as the vehicle starts moving, with riders
shouting out the “name” of their stop (usually a nearby landmark). With the
varied and precise fares there is a constant issue of having adequate change;
it is common practice for a minibus to pull into a petrol station mid-route in
order to simultaneously gas up and seek change. With the back seats of the bus
being beyond the reach of the conductor it means that the fare payments and any
change given need to be passed from hand-to-hand. The process is even more
complicated on the busy routes where large buses are used and money needs to be
passed over as many as 5 rows of intermediaries, creating plenty of
opportunities for dropped coins.

With the
operations focused on full and fast-moving (where possible) vehicles, “problem
riders” are not particularly welcome on Lusaka minibuses. Although it is not
uncommon to see a rider board with construction supplies or a massive bag of
corn flour (with the payment of an additional fare), I have yet to see a rider
with a disability, nor a parent with a stroller, sights which are commonplace
during my transit journeys around Canadian cities. Strollers are likely a moot point for most Lusakans as in Zambian culture the traditional way
for a parent to carry a child is using a chitenge with the
baby wrapped around the mother’s back. As for the disabled, the current situation at least disadvantages, if not outright excludes,
riders who have practical difficulties in meeting the expectation of boarding
quickly and squeezing oneself into a tight space.

minibuses is far more affordable than owning a vehicle or riding private taxis,
but still expensive on local terms. One advantage of a radial network in a
system where every transfer means a new fare is that almost every
origin-destination pair involves only one connection, and making the journey a
little cheaper in the process. I learned this myself in my daily commute for
language classes from my home in Roma (north part of the city) to Bauleni
(south-east). When I would travel there through Town there was only one
connection and the fare was 9.50 kwacha (currently 5.5 kwacha = $1US), and the
one-way trip would take nearly two hours with the traffic. Through some
experimentation I learned how to make the same journey using a combination of
inbound, outbound, and orbital routes (including a shared taxi), requiring five
connections, but dropping my travel time to under 90 minutes. However, the
total fare rose to 19.50 kwacha. The entire journey from home to class by
private taxi is about 25 minutes in the morning rush hour, but costs me 80
kwacha. Meanwhile, until quite recently the minimum monthly wage for a domestic
worker in Zambia was 480 kwacha for up to 24 days of work. Had a maid or
cleaner been making the same trip that I was (which is not implausible), using
the slowest and cheapest routing, the fares would have consumed the
equivalent of an
 entire month’s salary.

reality that travel is expensive helps to explain the ridership behaviour of
patiently waiting at stops while vehicles take an eternity to fill and depart:
a rider is always free to get up and leave the vehicle before their
destination, but by doing so (s)he will forfeit the fare paid early in the
ride. It is possible to chase down the conductor and try to persuade the
granting of a refund or a rebate, but this is a discussion in which a rider
holds precious little leverage. Indeed, there is a more severe form of this
situation where minibuses have maintenance issues; it is not uncommon to see a
crowd waiting near a vehicle that is jacked up for a tire change so that those
riders can benefit from the remainder of the trip for which they have already

Reflecting on the ride

foundational principles of this system ensure that it will only remain a
viable, let alone enticing, option for the primary transportation needs of a
certain section of the population: those in the income band that can afford to
use it but not afford greater convenience. Presently, that population is large
enough, and their expectations low enough, to support an entire industry. The
system exists at the convergence of a price point that is accessible to a large
number of riders, the profitability required for operators to earn a meagre
living, and a level of functionality where riding is (usually) at least a bit
preferable to walking. In this respect, public transportation in Lusaka is
similar to a large and growing number of rapidly expanding cities where the
current public policy is that the government offer the absolute minimum of
attention and interest. The rider’s experience on such a system is directly
proportional to that interest and attention.

Beyond income, there is
also an element of class association; some Lusakans that I
have spoken with do not consider minibus travel even for the trips where it
could make practical sense and in the situations where it definitely makes
economic sense. It seems to be a common sentiment in more elite groups that
staying home is preferable, “as that is not a transportation
option for me/us.”

minibus riders and operators not well-represented among the economically or
politically powerful, this form of transportation seems to be framed as a
problem (rather than a solution) in the dominant discourse. One seemingly
common framing is that of minibuses “creating congestion” by veering from their
designated routes or stopping outside of designated areas. With this framing
the apparent solution to the hot-button issue of congestion is the application
of increasing constraint to minibus operation. Similar views are common regarding
the enforcement of regular
departure times
, and the obligation that operators accept passengers with disabilities
(Persons with Disabilities Act). There have even been calls for citizen
surveillance regarding vehicle conduct
and maintenance
. Such strategies use the stick while neglecting the carrot: the current
realities of discomfort and unpredictability are in fact rational products of a
system where the prime incentivizing force is the ability of a multitude of
independent operators to generate a very thin margin of subsistence profit.
Although it is possible to improve the situation experienced by riders through
greater regulation and enforcement, these options should be recognized as one policy
stream among many; and that these options need not be limited to those that
shift the burdens from riders to operators.    

“MaConducta’, nisala!”: Disembarking

its pitfalls, public transportation in Lusaka is generally safe, as understood
from both crime and road safety dimensions. It is, however, intimidating to an
unfamiliar rider, and there are no mechanisms to encourage rider orientation at
a system-level (note that I do not consider having a conductor run up and grab
your arm while asking “Where are you going [right now]?” to be
reasonable orientation). This post will hopefully serve as a de-mystifying
agent for at least some potential riders with open minds, patience, and an
interest for a respectful taste of the daily reality of many Lusakans. May we
collectively establish how to maximize the system for what it does well while
systematically steering clear of its worst issues.

singapore: bbc profiles new frontiers in transit denial

This just in from the BBC:  Technology giant Philips corporation sent some people to the extremely busy Singapore bus system to imagine an alternative to typical fixed-route bus service.  The researchers' definition of the problem:

We discussed the benefits and limitations of the fixed-route system – it's clear such a system provided consistency in time and place (to get on and off), and to a certain extent convenience, but not completely. Flexibility is not what a fixed-route and fixed-time bus service system can offer. We have all experienced times when the bus is very empty or extremely packed, which means efficiency is best optimised at the bus-route level, but not individual bus level, since that bus is unable to respond to dynamic demand and traffic situations immediately. We all have all been in situations when there are only a few passengers in the bus and yet, the bus still has to plough through the entire fixed route, picking up no passengers along the way. The motivation was how to optimise the bus service by allowing the passengers and bus drivers to respond immediately to dynamic demand and traffic situations, not unlike a taxi that you can flag anywhere, anytime, and it will take you directly to your destination.

Needless to say, they came up with a massively all-demand-reponsive system identical to the one promoted last year by Gensler Associates, to which I responded (perhaps too colorfully) here. The idea is that now that you have a smartphone, the transit line should twist and turn to meet chase everyone's speciic need and that somehow this will be more efficient.  As I said in response to Gensler, there's little to fear from this dystopian vision beause it's mathematically impossible.  

In a place as crowded as Singapore, well-designed scheduled fixed routes are not just efficient but liberating.  They're efficient on a large scale despite routine under- and overcrowding because they follow straight paths that thousands of people find useful at the same time.  They're efficient because people gather at major stops where they board and alight in large numbers that are impossible in any demand-responsive form.  Frequent fixed routes are liberating because they're there for you when you need them, just as subways are, so that you don't have to wonder whether some automated system will approve your request for transport.  

The all-demand-responsive vision can mean one of two things:  (1) large buses that carry large numbers of people on complex variable routes, changing its route in response to every beep of desire from each of 5 million phones, or (2) fleets of very small vehicles each serving a few people on a more direct path.  Vision (1) is a hellishly circuitous system to ride any distance on, while (2) is a vision of vastly more wasteful use of urban space,  as people who are now carried in a space-efficient way are converted to a space-wasteful one.  Vision (2) also requires either driverless technology or extremely cheap labor, which is why it only happens at scale in low-wage developing countries.

No, Singapore has built its success on subways, and is developing fixed, infrastructural bus lines that work more like subways.

Please don't call yourself a transit visionary until you've grappled with the facts and possibilities of transit network design, by reading a book, say, or taking a course!

How urbanist visionaries can muck up transit

Architects and urban visionaries play an incredibly important role in a leadership-hungry culture.  They have to know a little bit about almost everything, which is hard to do.  But for some reason, certain segments of the profession have decided that the basic math and geometry of transit isn't one of those things they need to know, even when they present themselves as transit experts.

To see what I mean, I encourage you to watch this short video from Gensler Architects in Los Angeles.  It's a concise summary of all the crucial mistakes that you'll need to confront in much "visionary thinking" about transit.  (If Gensler takes down the video, read on.  I've inserted enough screenshots from it that you can follow.)


[NETWORK_LA transit from tam thien tran on Vimeo.]

The five most common "visionary" mistakes about transit, all on display in the video, are:

  • Disinterest in costs and efficiency.   Visionaries do need to set aside cost and efficiency for part of their brainstorming phase, because by doing so they might come upon an idea that's efficient and affordable in a completely new way.  But they don't have a coherent idea until they've brought those factors back in, at least at the level of order-of-magnitude reasonableness. Sadly, some urbanists scoff when I use the word efficiency, assuming that this means I've lost touch with human needs, aspirations, aesthetics and values.  In reality, efficiency means how much of those good things you can have in a world of limited resources.  Even in the arts, we speak often of the efficiency or economy with which an artist achieves an aesthetic effect.  (The Gensler video, for example, is efficient in displaying all five of these fallacies in only five minutes.)
  • Fixation on transit technologies as though they were the essential distinction between different  mobility outcomes.  For more on this, see here.
  • Confusion about scale.  In transit, if it doesn't scale, it doesn't matter.  Because visionary thinking often focuses first on a prototype – a tiny example of the hoped-for transformation — it often goes too far without thinking about scalability.  Sure, this cool idea works in one suburb or in one cool building, but that says very little about whether it would work in a whole city.  Gensler's particular error about scale is … 
  • Confusion about "flexibility," a dangerous slippery word.  Gensler imagines that a demand-responsive style of transit, in which you make a request on your phone and the transit system somehow deviates to meet your personal needs, is scalable to a vast, dense city where the transit system is already very crowded much of the time.  More on this below. 
  • Ignorance about what's already working, leading to premature demolition fantasies.  If you already hate buses, you won't have much interest in understanding why so many people use them.  Like many urbanist visionaries, Gensler doesn't appreciate the very high ridership and efficiency of the existing transit system across the core of Los Angeles. This allows them to jump to the conclusion that the system should be replaced instead of incrementally improved.  (Tip:  Prematurely dismissing the relevance of something that so many people clearly find useful is an excellent way to sound elitistregardless of the nobility of your intentions.)

So watch the Gensler video if you can, but you can also follow along via my screenshots and comments below.  You'll see these mistakes again and again in the urban visioning business.

0:27 Gensler states the question as "Get LA on transit HOW?"  No argument with the question.


0:51  Transit is divided into a set of vehicle types, and these types (light rail, metro, bus) are confused with "methods" of transport.  For more on the absurdity of treating bus/rail distinctions as primary, see here.


0:53  "We have only these methods.  What if we added more?"  An interesting question to which transit experts (and economists, and engineers) have a very good answer.  The more competing systems you establish in the same market trying to do the same thing, the less well any of them will function, and the less investment any one of them will justify.




0:56  They now begin to analyze vehicles in terms of distance, sustainability, flexibility.  What's missing?   Cost!  Efficiency!  Some things are just wildly expensive relative to what they deliver.  Darrin Nordahl has already been down this path, evaluating technologies by discussing only their supposed benefits.  That's not evaluation, it's either aesthetic rumination or marketing.  (Neither of those are bad things, but they have to be identified as what they are.)


1:20.  They talk about distances but their graphic is talking about speeds.  These are fair for personal modes but absurd generalizations for the transit modes. When your notion of "rail" conflates light rail, heavy metro rail subways, and 70 mile-long infrequent commuter rail, the word "rail" means nothing relevant about speed or travel distance, or any other transit outcome apart from capacity.  (Note that the earlier claim "we have only these methods" implies that these three kinds of rail are the same thing in every way that matters.) 

Likewise, if you think buses have an ideal distance, you're unclear on the role of local buses vs Bus Rapid Transit vs long-haul expresses, all of which are very successful in Los Angeles.  Gensler imposes a "technology first" frame on the data, thereby concealing almost everything that matters about how transit gets people where they're going.

In transit, the real speed distinctions within transit are usually not direct results of technology.  Speed is the result of how often you stop and what can get in your way.  See here.





2:00.  Staggering incoherence in comparing input (bus service) to an unrelated output (total ridership including rail).  What's more, the numbers are misleading.  Per the 2011 APTA Fact Book, Los Angeles MTA has America's 3rd highest total boardings and 2nd highest total bus boardings.   In the context of its starved resources and the vagueness of public support for it, the Los Angeles bus system is working brilliantly.

2:26.  Here is Gensler's biggest mistake:

Gensler 1

Gensler 2

Which of these two networks would you rather travel on?

Gensler has mistaken metaphor for logic.  They think that "liberating" bus routes has something to do with liberating or enabling people.  The idea is barely explained and totally incoherent. 

Today, in our supposedly "inflexible" system, you'll find a bus going down a major boulevard with maybe 60 people on it.  Some of them want to go somewhere straight ahead, some want to go to somewhere ahead and to the left, some want to to somewhere ahead and to the right.  Fortunately, they are in a high frequency grid system, which will take all of them to their destination, either directly or via a connection to a north-south line, probably by a path similar to what they'd have followed if driving.  So this huge number of diverse people making diverse trips are all moving toward their destinations on a reasonably direct path.  This is the extraordinary power of the high-frequency grid.  So instead, Gensler proposes bus lines should twist and turn just because somebody with an iPhone wants them to?

Personal technology has great opportunity to better inform us about all transit services, and it can transform the convenience of transit at low-demand places and times, by influencing the operations of low-ridership, low-capacity services, such as demand-responsive buses and taxis. 

Quite possibly, personal apps will allow demand-responsive service to replace some low-demand fixed-route buses, which is fine with most transit planners.  Those low-ridership buses run mostly for social-service or "equity" reasons, and if there's a more efficient way to do that, I expect many transit experts would be all for it.  It would let them concentrate on the high-ridership, high-capacity services that can achieve a great deal of personal mobility and sustainability, very efficiently. 

Successful high-capacity frequent transit needs to take on more of the rigidity of subways, in order to spread the benefits of subways (which we can't afford everywhere) more widely.  That means it needs to be even more frequent, reliable, legible, permanent, and reinforced with infrastructure investment.  Fortunately, within limited resources, many transit agencies are now trying to do that.

The video is full of entirely laudable and familiar green ideas, but then we get to this …

  • 3:23  In Gensler's Los Angeles, every transit trip must be reserved.  Do you really want to have to make an appointment with a single vehicle and driver, because that's the only way to make any use of all the buses swarming around you on unpredictable paths?  Or might you prefer a simple frequent transit corridor where so many buses are coming all the time, in such a predictable pattern, that you can take any of them, and are thus almost guaranteed a vehicle soon even if one breaks down?


  • 4:20  "What if we had PERSONAL service?" they ask?  Well, the extreme of personal service would be low-ridership system in a tiny town, where the driver has time to learn everyone's name.  Is that what Los Angeles wants to be?   Or would you rather live in a city where you can get anywhere you want to go easily, starting right now, without making a reservation, and even with the option of spontaneously changing your path or destination, just like motorists do?  

To me as someone who values my personal freedom, flexibility, spontaneity, human dignity, and travel time, Gensler's Los Angeles would be a hell-world worse than Blade Runner.  Fortunately, it's also mathematically impossible.

We've blown up transit networks before, of course, and Gensler's vision should remind us of what was thought about cars vs. transit in the 1940s.  Like personal technology today, cars were just so wonderful for the individual that we just assumed the world could be made in their image.  (The technical term for this idea — that the world will bend to reflect my emotional needs and enthusiasms — is narcissism.)  So we made a deep investment in a car-and-highway technology that could not possibly scale to big cities.  Gensler proposes the same mistake:  Because our iPhones are so cool, they assume that the city, at every scale, can be reinvented around them.

For a more positive vision of the future of Los Angeles, one that begins by noticing the city's strengths and looking at how to build on them, see here and especially toward the end of an interview here.


Can Dial-a-Ride Get High Ridership from Low Density?

In this post, I argued transit can’t be judged on the low ridership of
services where ridership isn’t the goal, and explained that every transit system has “Coverage” services, designed to achieve a perception of equity and/or to meet the severe needs of small numbers of people.  Coverage services generally cover low-density areas where ridership will always be relatively low.

In the comments, David Marcus asked a really important clarifying question, one that I hear often from elected officials in low-density places:

Could the empty-running coverage-oriented buses be replaced with some sort of dial-a-ride system running full?

Continue Reading →