More on My Light Rail Bet

Thanks to Tom Kirkendall for the link to my light rail post.  For quite a while, he has been "railing" against Houston's light rail proposals (where I was born and raised).  By the way, he is right that Phoenix is even less amenable to a rail-based system than Houston.  Houston has low population density and its downtown area is small compared to metro-friendly cities like New York, making rail an iffy proposition.  But Phoenix is even less dense and its downtown is tiny compared even to Houston.

A previous post of Tom's also gives me data to feel even more confident about my proposed bet, which was this:

If we take the entire cost of the system's construction, plus its
annual operating losses/subsides, I will bet that we could have bought
every regular rider of the rail system a nice car instead and gas for
life cheaper than the cost of the rail system.

Obviously we don't have Phoenix numbers yet, but he links an LA Times story with Los Angeles numbers:

Three light-rail lines have been added to L.A. county's transit system
in the last 20 years. Together, these cost $2.5 billion in capital
costs, they serve about 125,000 passengers per day and account for a
fiscal loss of approximately $252 million per year -- if one
acknowledges that capital costs are real, something that transit
operators and boosters often neglect.

Note that LA's system is actually a more desirable system from a rider standpoint than the one in Phoenix, since in some areas the trains avoid traffic lights, making them closer to heavy rail, and thus have a faster speed.  So lets run my bet against LA's numbers.  We don't really know what the core ridership numbers are.  Certainly its less than the 125,000.  And we don't know if an out in the morning and back at night commute counts in these numbers as one passenger or two (From here, it looks like 125,000 passengers making 2 trips each).

If the core ridership number is 125,000, the highest possible choice, then the total capital cost of the system per rider is $20,000 per rider.  This means I was right, that we could have instead bought ever rider a car for the same money.  Since the real ridership is probably less than that number, this means we could have bought ever rider a car and had money left over.  Concerned about the environment?  Then make every car a Prius, which the money would just about cover even without the volume purchasing discount they would likely get.

But what about gas?  Well, they say they have a $252 million per year operating loss.  This subsidy, which is above and beyond ticket sales, equates to $2,106 (!) per daily rider, even using the higher 125,000 figure.  At $2.50 per gallon, this equates to 15.5 gallons of gas per rider per week. 

So you can see with the LA numbers, even using the largest possible interpretation of their ridership numbers, the money used for the train could have instead bought every passenger a new car and filled the tank up with gas once a week for life.

Yes, I know, the argument is that the train reduces congestion.  Supposedly.  I have two responses:

  • Rail has never reduced congestion in any city.  Go see London and Manhattan.  In fact, rail seems to encourage urban density that increases congestion. 
  • In Phoenix, where rail will often replace existing lanes of roads, the train will likely carry fewer people than the lanes of traffic used to, so congestion will increase.


  1. Rob:

    Light rail sucks in San Jose.

    -Ridership is low.
    -It screws up surface streets - they couldn't be bothered to elevate, excavate the rails or suitably adjust the streets. They share and that sucks.
    -It is slow relative to auto travel (see above about streets). If you want me to use it, it has to superior to my current solution, a car. I used it now and then, when I was attending San Jose State and worked along the line. This is when I learned that I could regularly "beat" the train when going to school.
    -Schedules suck. The subway is useful because the wait time between trains is not bad. This is due to decent ridership - which we don't have.
    -IT DOESN'T PAY FOR ITSELF. I should have said this first and left it at that.

    Again, municipal rail systems suck without high ridership, which is engendered by high population density. It is something for bored transit and urban planner folks to do.

  2. Matthew Brown:

    In LA, in fact, the three light rail lines are largely on dedicated rights-of-way; the Green Line has a dedicated right of way built into the center of the 105 freeway for most of its length and an elevated right-of-way for the remainder of its run.

    (Incidentally, much of the point of this line was lost when the MTA caved to political pressure and did not connect the Green Line with LAX airport directly. The pressure was from airport parking, car rental and taxi interests. Tourists and visitors are a constituency most likely to use public transport if it's good enough. It also did not connect with the I-5 and the Metrolink trains on the other end of its run as originally envisaged as well)

    The Blue Line runs on streets at both ends of its run (Downtown LA and Long Beach, respectively) but runs on what was once the Pacific Electric dedicated right-of-way for most of the rest of the run. The Red Line also has its own dedicated right-of-way, underground for the first half. The Gold Line to Pasadena runs on railroad trackage as well.

    Thus Los Angeles experience is not comparable at all with street-running light rail, and is a more desirable (albeit, I'm sure, more expensive) system.

    LA also has an extensive urban population who use transit extensively. It's still notable that bus riders, the Bus Riders' Union, sued the MTA to halt rail construction and use the money on less expensive and more widely-used bus service improvements.

  3. Matthew Brown:

    I should point out that I'm a hardcore railfan and I still don't see the point of most light rail implementations. The fact is that American cities once had extensive urban and suburban rail service and they mostly closed down because once the car became affordable, nobody used the rail service. Sure, people mourned the closure of Pacific Electric or the Key System, but the fact is that they did not ride them - they wanted them around, but not on their dime. Urban and interurban electric railways seemed like a good idea when the competition was walking, horses and full-size steam railroads, but against cars there was no contest.

    Rubber-tired buses are also preferable; cheaper, better economies of scale, need less capital investment, can reroute around obstacles, and can go anywhere. Even dedicated busways make more sense than light rail, although I'm not convinced the capital investment makes sense there either.

  4. Ironman:

    One of the places that you would expect rail to work for commuting would be up in Seattle - here, you have a well-defined downtown area, high density population and a very "pro-mass transit" population, at least as measured by how they vote for light-rail and mass transit initiatives.

    In practice, it's a near complete failure - ridership is nowhere near the projections used to justify the expense, and even with costs sunk, it's nowhere near operational profitability. Economically speaking, of all the ways to transport people between cities, rail is perhaps the stupidest. Nowhere else do we see the confluence of extraordinarily high infrastructure costs (land, construction, equipment, facilities, etc.) and extraordinarily high operating costs (labor, maintenance, fuel, utilities, overhead, etc.) combine with extraordinarily low demand by commuters to produce such little tangible benefit. The bottom line: rail sucks.

    Just a thought - once you get the numbers (and win the bet), maybe you could use the proceeds to sponsor this mass transit initiative....

  5. murraymises:

    Check out this report from the Federal Reserve Bank in St. Louis. It estimates “the annual light-rail subsidies [in St. Louis] could instead be used to buy an environmentally friendly hybrid Toyota Prius every five years for each poor rider and even to pay annual maintenance costs of $6,000.” Not just any car, a Toyota Prius.

  6. Matthew:

    I am particularly enjoying your analysis of this issue, and would appreciate it if you could look at some of the other alleged selling points of light rail, such as reduced pollution.

  7. Zoran Lazarevic:

    New York City subway actually decreased the density of the city. Manhattan had the highest population in 1904 when the subway opened and allowed working poor to migrate to the Bronx, and later Queens and Brooklyn.

    However, I still use a car when going downtown in the evenings and weekends. Trying to use a car on a workday 9-5 is usually slow and expensive for parking.

  8. JohnDewey:

    A few points to consider about light rail in general:

    1. Many light rail riders - as much as a third in Dallas - were riding buses without complaint before the trains started. Their use of trains did nothing to reduce congestion or pollution.

    2. Many bus riders saw increased commute times after trains started, as bus routes were rerouted to feed trains and increase reported train ridership.

    3. Transit authorities count each one-way, unlinked trip as an additional rider, intentionally obscuring the fact that the number of people using trains is less than half their reported ridership.

    4. Free parking at transit stations leads many car commuters to utilize teh last few stops as parking garages to avoid downtown parking fees. These "riders" are really car commuters whose parking costs are subsidized by taxpayers.

    5. Pollution from heavy equipment used to build the rail lines will exceed many years of the reduction in commuter pollution.

    6. Congestion around transit stations increases the overall pollution in a city. So does the traffic stoppage and traffic rerouting during both construction of tracks and operation of trains. Together with the heavy equipment increase in 5. above, the added pollution likely exceeds any "green" benefits of light rail.

  9. Alexander Holt:

    There is one that I think you got wrong in your analysis. The system had $2.5 billion in capital costs. That number directly impacts the $252 million dollar year operating lost (amoritizing the loan/bonds). Then you say that you could buy all the riders (125,000) a Prius or whatever with that $2.5 Billion and pay for the gasoline yearly with the $252 million. But that using the $2.5 billion twice.