Transit in Phoenix

A reader writes
Our AZ Senate candidate is very big on light rail and says in debates that it is a good option for our District – NW Phoenix. He thinks ‘commuter rail’ has been effective in Chicago and wants to make a case for light or commuter rail here.
From your perspective, are either methods successful in cities in the US? One measure of success is in terms of reducing some measurable amount of automobile traffic [probably 5% or more] or a different measure is in terms of public investment – such as making the system pay for itself after some period of initial investment.
Commuter and light rail are completely different beasts, like comparing taxis and buses, they both move people, but they move different numbers of people at different speeds for different distances.
Defining success in terms of reducing auto traffic is also a mistake, recalling the Onion headline Report: 98 Percent Of U.S. Commuters Favor Public Transportation For Others (November 29, 2000 | Issue 36•43)
We don’t define airplanes in how many rail passengers they take off the train, or the success of typewriters in how many word processing users avoid computers.
So there are a few measures we might consider. A private firm would ask: Do the marginal private benefits (profit) outweigh the marginal private costs? This is how transit projects were judged back in the day (the late 1800s and early 1900s) when they were private. By marginal we mean does the next dollar invested have benefits that outweigh the cost of one dollar.
Alternatively, does rail provide transportation for their users at a cost they pay for? The answer is clearly no in every US city.
As far as I know, passenger rail now only makes money in Hong Kong. It probably could break even with appropriate management in a few other cities (e.g. New York, London, Paris), but everywhere else it is heavily subsidized (100% of capital costs and two-thirds of operating costs is typical subsidy for rail transit in the US).
We might distinguish “does pay for” and “would be willing to pay for” and consider the notion of subsidies, but I doubt any system (outside New York and a few select routes elsewhere) could break even at their existing costs.
The public asks do the marginal social benefits (MSB) outweigh the marginal social costs. Marginal social benefits in theory might include non-user benefits like congestion reduction, pollution reduction, crash reduction, noise reduction, increased accessibility for non-users, and so on (to the extent these can be accurately measured and monetized) but MSB would primarily be comprised of user benefits (those accruing to the transit riders themselves). The marginal social costs (MSC) are the “private” costs of paying for the infrastructure and service, and any externalities that are created (delay during construction, pollution caused, crashes caused, delay to non-users during operation, etc.).
The monetization of some of these costs depends on personal values, value of time, value of life, value of health, value of quiet, and so on, though economists and engineers have assigned values to these (value of time =$10/hour, value of life =$3 million, …) based on individual choices when making real decisions.
However, we also need to consider the alternative use of resources, the opportunity cost. If we spend money on X we no longer have it to spend on Y. So even if X is good, Y might be better, and resources are scarce.
Again I believe the answer is no, the marginal social benefits seldom outweigh the marginal social costs in fixed-rail transit investments. I have not seen any benefit cost analysis that I believe that has a rail project with marginal social benefits exceeding marginal social costs.
Rail advocates then claim there are non-monetizable factors, civic pride or image, etc. I remain unconvinced.
Alternatively, they may suggest that the values of things are underestimated (e.g. economic development, land use changes), but usually this double-counts benefits that are in the analysis. (land values plus user time, e.g., by and large are representing the same thing, the reason land values are high is because the land is located in a place with better access (less time to more things)).
A final argument concerns environmental benefits. In general cars pollute more than electric trains (though this depends in large part where the electricity for your light rail comes from). But the value of this can be monetized w health expenses. (Or worse, diesel for your trains may be more damaging than auto emissions.) The cost of global warming is another matter which is highly speculative.
Phoenix’s current transit (bus + rail) work trip mode share is 1.9% according to the 2000 Census Transportation Planning Package data. To take 5% more cars off the road, transit mode share would need to be more than 7% (since some of those travelers would come out of carpools, walking, and biking, and non-work mode-share is less than work trip mode share). The problem is compounded by the idea of induced demand, by reducing congestion, some people who previously avoided traveling at peak times would now travel then again (for every 100 trips removed because of transit, maybe 50 or so would then be made, changing time of day or day of week, making longer trips, switching from carpool to drive alone, or switching from other routes, or making trips previously avoided).
The only large US Cities with a 7% or higher transit work trip mode share: New York, Chicago, Washington DC/Baltimore, San Francisco, Philadelphia, Boston. Large portions of these cities were built in the transit era and they have all had long-standing transit systems. Phoenix is not likely to quadruple transit usage without a very large investment (transit service as extensive or more than the cities identified above … I say more because the land use pattern in Phoenix is much less conducive to transit than the cities above) … unless there is a large external shock (e.g. very high gas prices, maybe on the order of $10 gallon).
Phoenix on the other hand has one of the highest carpool share in the US. This may be due to HOV network, but is more likely because of a high number of working class individuals who are sharing cars to get to work. Exploiting this predisposition to carpool seems more promising than trying to jump-start a new mode.
If you want to reduce congestion, you have to increase prices, ideally prices that are targeted by time of day and location, to give people the appropriate signals about the real cost of their travel. If you are unwilling to do this, your congestion is not bad enough.
— dml