A collection of essays by Bill (website@ccjj.info) accompanied by feedback from his friends.

Monday, December 28, 2009

Suburban Public Transport

Description of Problem

A large share of the residential neighborhoods of the United States consist of suburbs, which have very low population density, consisting of one or two story residential buildings with lawns. Among the adult residents of these neighborhoods, car ownership is nearly unanimous, public transportation in these areas being of extremely poor quality, consisting of routes that are very far apart, with vehicles that come extremely infrequently, and most of these public transport systems are utterly dependent upon government assistance to exist since they are completely unprofitable.

Having every adult in the population own and drive a car is not a sustainable situation for the coming decades. It involves a high use of energy, and the preferred form of energy used for operating automobiles is the combustion of liquid fossil fuels, which 1) has negative implications for global warming, 2) has negative implications for smog, 3) will run into serious problems as oil reserves get depleted just as the developing world increases its appetite for these fuels, 4) involves a dependency upon a few countries that, in recent years at least, the United States has not been getting along with very well.

It is possible that the US could be in for a rude awakening, due to some sort of political change such as a nuclear-armed Iran taking over the Persian Gulf, or the collapse of the House of Saud, sending oil prices to hundreds of dollars per barrel and many Americans suddenly finding it cost-ineffective to get to work. We need to change our transportation system, and the sooner the better.

In some American cities such as New York, particularly Manhattan, population density is high enough that high-quality public transport is possible (though it still needs government subsidies). However, as energy and environmental issues become more pressing, it would be extremely economically painful to massively move the population into cities, abandoning a fortune in suburban real estate. Since the peak in 2006, real estate prices have dropped 20% in this country, and that drop is destroying our financial system. If all the suburban real estate were to be abandoned, thereby losing 95% of its value, the economic consequences would be much more dire.

Unacceptability of Current Alternatives to Passenger-Owned Cars

In a typical American suburb, a consumer has to walk a very long way to get to a bus stop going in a desired direction, and then wait a very long time for the next bus to arrive. Rarely is the consumer lucky enough that a single bus route stops both at the start and end of his trip, he has to take multiple buses, waiting something like a half hour for each one.

With trains the problem of buses is exacerbated since special tracks are required, and a train is as big as many buses, so the tracks are even further apart than bus lines and the trains come less frequently.

With taxis, the expense of the driver is very great for a given amount of transportation, even though the driver is often working ridiculously long hours for very little pay. In addition, cities often severely limit the number of taxi licenses issued, resulting in major areas (like all burroughs of New York except Manhattan) being completely underserved. In a place like Manhattan, a consumer can hail a cab from the curb (at least on the right street, in the right neighborhood). In a suburb, that option is totally unavailable, one must call a cab company and wait for a cab to come pick one up, which typically takes a long time. Taxis are generally an extremely expensive way to get around.

It is generally understood by American suburban dwellers that it is so difficult to survive without a car that only really, really poor people even try. People who've had their licenses revoked usually drive illegally, people who can't afford insurance drive without it.

Part of the reason public transport is so bad in suburbia is that almost no one is using it, making the effective population density of actual public transport riders much, much lower that the density of the total population. If an acceptable public transport system could draw a critical mass of customers, the response time of the service would improve, resulting in a virtuous cycle of "improved service begets increased ridership begets improved service ...".

Desirability of Larger Vehicles

Larger vehicles are more economic and energy efficient than smaller ones. At highway speeds, most of the energy expended by a car is in the form of air friction. A bus has less air friction than enough cars to carry all the passengers in it, and the air friction a train encounters is negligible compared with that encountered by enough cars to carry its load. Also, larger vehicles need fewer drivers per rider carried. The technology for large transit vehicles exists and is well-developed.

The Last Mile Problem

It would be nice to be able to get everybody in large efficient vehicles to get efficiently to their destination, but these large vehicles can't stop at each rider's start and destination. Getting everybody that "last mile" to and from their thinly spread-out homes is key -- if we can solve that problem, we can then get people within reach of trunk lines of efficient, large-scale transportation that we know how to do well. Solving the "last mile problem" is key.

Jitneys or Shared Taxis

In the United States, "jitneys", also known as shared taxis, that is cabs carrying several independent passengers, were quite popular in the early 20th century, but were mostly banned due to pressure from public transportation monopolies. In many places, particularly the third world, shared cabs are today quite popular. Due to low wages in the third world, the cost of the driver is low, and due to the vehicles being small, the routes are close together and vehicles come frequently, providing very good service at a low price. More on Share Taxis

Computer-Routed Jitneys

Computer navigation systems are now quite common in cars, the technology is very well-developed. It should be possible for a rider to call a dispatching service that knows the location of all of a fleet of jitneys, which would issue orders to one of the jitneys that is near the rider and has a spare seat to deviate from its route to pick up the rider and take them toward their destination. The jitneys would follow no fixed routes, but roam about the suburb following orders from the dispatching service. The jitneys could go directly to addresses rather than riders having to congregate at stops. The driver would not know the details of where the jitney is going to go, he would just follow short-term instructions from the navigation system which is being driven by the dispatching computer.

Requesting a ride will be particularly easy because so much of the population already has cell phones, and most cell phones have GPS receivers in them, which could be incorporated into the system, so a rider would call the dispatcher and the dispatcher could instantly see from the GPS where the rider was, then needing to find out only where the rider wanted to go. More advanced phones such as iPhones and Blackberries could have efficient interfaces for requesting rides to favorite destinations, simplifying the process and cutting down on the expense of operators. Riders could also request rides from computers at home or at work.

Multiple-Vehicle Routing

There is no reason one would have to spend an entire trip in a single vehicle. You could be picked up near your home by a jitney, dropped off at a train station or bus stop, then ride the train or bus express 40 miles at high speed with few stops, then take a jitney the last 2 miles to your destination. So most of your trip is done efficiently, at high speed, with fewer stops.

If you're not in a hurry, you could get a cheaper cost ride for short distances by taking multiple jitneys. For an oversimplified example, if you were going about 6 miles northwest and most of the traffic in your neighborhood was east-west and north-south, you could catch one jitney 4 miles west, be dropped at a corner, and picked up by another jitney headed north which will take you to your destination.

If a large amount of the traffic were to eventually shift from driving their own cars to riding this public transit system, we could find ourselves in a situation where most of the freeway traffic is people riding in buses and jitneys, resulting in far fewer vehicles on the freeway for less congestion.

Computer-Driven Vehicles

Because a jitney carries many more passengers than a taxicab but many fewer than a bus, the cost of the driver is less problematic than for a taxi driver, but still much more problematic than for a bus or a train driver. If the jitney drivers are paid as poorly as NYC cab drivers, who I am told work 72 hour weeks, cheat on their taxes, and then take home only $20,000 a year, well, that's just gross. If they are paid as well as NYC subway drivers, who are paid $55,000 a year and retire with half-pay at 55 years old, the cost of the jitney service would be prohibitive.

The technology for computer-driven vehicles is getting fairly mature. In 2007, DARPA held an event called the Urban Challenge where university teams built unmanned, robot-driven autonomous vehicles driving a through streets in a neighborhood (actually an abandoned military base) with other traffic and traffic signals. Six vehicles successfully completed the course. Computer-driven jitneys could be an extremely cheap way to get riders around. When traffic is limited, jitneys could just park somewhere, turn off the engine, and wait for someone to want a ride. It could still take a few years before computer drivers are good enough that we will want to trust them driving around suburban neighborhoods with children and pedestrians in the streets.

Evolution of the System and Political Considerations

Evolution, not Revolution

Generally it is vastly preferable, with any new idea, to start small, prove the concept, learn from experience, and grow, rather than make a sudden change. It is unnecessary to plan a scenario where everyone junks their cars and starts riding jitneys overnight. Some people will really like their cars and want to continue using them. Also, raising initial funding a system capable of assuming the entire load of a suburb would be a huge problem.

The biggest reason that people have resistance to this whole idea is that they think I am proposing that everyone will abandon their cars and rely 100% on computer-routed jitneys overnight, and that if that doesn't happen, the entire idea is a failure. This idea is a success if one can make a computer-routed jitney company profitable, and my assertion is that such a company could be profitable anywhere a taxi company is profitable, which, right now, is basically everywhere. If one has a company with 30 jitneys operating in a suburb sprawl of a million people, it can move 0.001% of the people there more cheaply than a taxi company with 30 cabs, and make a profit. It will, at that point, be able to sustain itself and grow to carrying an increasing proportion of the area's total transit load.

No Monopolies

There is no reason the jitney service has to be given a monopoly, We have benefited from allowing UPS and FedEx to compete with the post office, we have multiple cellphone companies operating in the same areas, we have multiple cable TV and internet companies serving the same neighborhoods, we have multiple long-distance bus lines in this country. Competition will serve the customer better than imposing a monopoly. Some services might specialize in providing the cheapest service, while others might provide a higher-quality of service, perhaps ensuring faster delivery by taking fewer riders per vehicle, or providing more comfortable seating.

Customers could also negotiate different price schedules -- for a higher price, the computer will place a higher priority on assigning a jitney to deviate to pick you up, and put fewer co-travelers on that jitney since stopping to pick them up / drop them off will slow you down on your way. People could thus gauge how much of a hurry they're in and decide whether to save time or money on a given trip.

Proof of Concept

Government is mostly an obstacle to this change. It was government that killed the jitneys in the early 20th century, many public transport systems have legal monopolies, and to implement this idea nationwide will cost billions of dollars of legal fees. The main strategy should be to prove the concept in areas where regulation is lax to nonexistent, and then other localities will want the benefit of the service and that will drive the necessary political and legal change.

We have lots of large suburban areas in this country. To prove the concept of the computer-dispatch jitney, it would be best to find an area that does not have a public transportation monopoly that is going to provide legal barriers to entry. New York City is out of the question.

One ideal way to get a foothold is doing "para transit". In many cities the government provides para transit, a subsidized taxi service for people who are medically unable to drive, such as blind people or epileptics. Computerized jitneys could enormously enhance the quality of service provided by these services, while providing developers a chance to get the bugs out of the system before they have to achieve the efficiency necessary for profit-making competition. In New York, an eligible rider can get a para transit ride for $2 each way, but they have to reserve a day or two in advance. Computer routing could greatly reduce the lead time between making a reservation and being picked up.

Once the technology is fully developed and the concept has been proven, I anticipate it will spread to other markets. For example, in NYC, taxis have a legal monopoly while, for stupid political reasons, the number of permits for cabs is kept so low that only Manhattan and the airports are serviced, leaving Brooklyn, Queens and the Bronx virtually without cabs. I think once the effectiveness of computerized jitneys has been established in other cities in the country, there will be no way to prevent the voters of the outer burroughs from voting to allow this sort of service in their neighborhoods.

Early Adopters

One argument I hear against this idea is that people like their cars. They aren't going to give them up easily. The answer to that is simple -- 150 years ago, people really liked their horses, and many people preferred animals to machines. But in the end, nearly everyone gave up their horses, not all at once, but slowly they did.

I feel that while many people absolutely love their cars, for many people owning a car is a big hassle. Many people, especially women, feel helpless taking their car to a mechanic because they fear being taken advantage of. And it's not a groundless fear. The median income in the US is about $43,000, with 20% of the population earning less than $18,000. Brand new, the cheapest cars sell for about $12,000 plus sales tax. Many people are afraid to buy used cars because they are not competent to deal with problems a used car may have, so buying and insuring a car is an enormous expense for a lot of people. If a cheap public transport alternative were available, many people will jump at the chance to use it.

Other than the poor, there are people who want to go out drinking sometimes yet are responsible enough not to want to drive home drunk. People medically unfit for driving, the blind, the elderly, epileptics, would all benefit greatly from this service, as would people too young to drive.

Resistance to Computer Drivers

There are a lot of people who drive vehicles for a living, and many of them are in unions, meaning they can organize themselves politically very easily. They will fight the introduction of computer driven vehicles to our roadways with everything they have. Another problem is that while computers won't make some mistakes human drivers will (for example, they will never drive drunk), they will have accidents of a sort that human drivers generally won't have. Lives will be lost, and it will be difficult to sell that to the public as an acceptable sacrifice. For example, the BART subway system in California originally had robot drivers in the '70's, until a robot malfunctioned, speeding a train up when it should have been slowing it down so that it crashed through the barriers at the end of the line. Ever since, BART trains all have drivers.


  1. This comment has been removed by a blog administrator.

  2. Good economist article: http://www.economist.com/node/21560989

  3. 7 years later, Scientific American is just starting to figure out the implications of driverless cars: http://www.scientificamerican.com/article/will-robo-ubers-kill-car-ownership/