Costs of Subway Slowdown Would Add Up Fast

Following the recent deaths of two subway passengers who were pushed onto tracks, TWU Local 100 is urging operators to slash train speeds as they enter stations, the New York Times reported yesterday. A TWU flier, which you can view here, advises operators that “Preventing a [run-over], and saving yourself the emotional trauma and potential loss of income that go with it, is worth a few extra minutes on your trip.”

There’s no question that watching death unfold through the train windshield and being powerless to avert it can result in trauma and guilt. But the proposed remedy in the TWU flier could be surprisingly costly. Based on calculations from my Balanced Transportation Analyzer spreadsheet model [PDF], if those “extra few minutes” were actually applied as a preventive measure to every subway trip, the lost time could aggregate to millions of hours per year for straphangers, not to mention more street and highway gridlock as the slowdown leads some commuters to drive instead of taking the train.

The city’s subways account for 1.6-1.7 billion passenger-trips a year. Here’s a rough sketch of the leading consequences from slowing all of them by an average of five percent:

  • A 2.4 percent drop in subway ridership, as slower service discourages “marginal” train trips
  • A nearly 4 percent rise in private auto trips into the Manhattan Central Business District, causing a 4 percent drop in average vehicle speeds there
  • 1 percent fewer people coming to the CBD — a net decrease of 34,000 each day
  • 55 million hours a year sacrificed to slower travel (35 million for transit users, 20 million for vehicle users), collectively costing them $1 billion a year, based on values of travel time
  • A $60 million a year revenue hit to NYC Transit, or a $35 million net loss for the MTA after factoring in higher throughput on tolled bridges and tunnels

These figures do not reflect higher personnel and equipment costs to run additional trains to make up for the slowdown. Nor do they capture macro-economic effects of reduced business from the decline in CBD activity. Even so, they’re not chicken-feed.

Analytically, the key driver of these impacts is an assumed “time-elasticity” for transit trips of negative 0.5. This means that increases in trip duration translate into half-as-great decreases in the volume of trips. Thus, an assumed 5 percent lengthening of the average subway trip (which is how I represented the “extra few minutes” urged in the TWU leaflet) would lead to a 2.5 percent shrinkage in subway use.  (The shrinkage settles out at 2.4 percent because the increase in CBD gridlock eventually “attracts back” a few of the trips that leave the subways due to slower service.)

To be sure, not every “lost” subway trip resurfaces as a car trip. Some non-work journeys simply disappear. Carpooling, though paltry, also cuts into a 1-for-1 substitution. So do buses and cabs, though the worsening of traffic due to greater auto use leads to predicted overall losses to both modes, albeit small. Even cycling might not grow, for the same reason. Still, the scale of subway use is so vast that even with these caveats, a minor (2.4 percent) decline still precipitates a proportionally greater (3.9 percent) increase in car trips to the CBD.

It’s not clear how many of the approximately 50 annual average subway-track deaths could be averted through a systemic subway slowdown anyway. Most of the deaths are suicides, and there’s little reason to think that fewer people will jump into the path of a 10 mph train than a 30 mph one. Still, transit operators deserve better than the three days off they’re granted after witnessing a death, not to mention the absence of any time off for most of the 100 other “minor” cases each year, as the Times reports.

Think of it this way: If each of the 150 track incidents a year led to one week off for one train operator, the cost to NYC Transit would be $300,000 a year (based roughly on fully loaded salaries of $100,000 a year). That’s 99 percent less than the $35 million I estimate a 5 percent slowdown would cost the authority in lost fares alone.

Subway track incidents need lasting and humane solutions that don’t undermine the efficiency of the city’s vital mass transit system. An across-the-board slowdown in train operation doesn’t appear to be one of them.

  • Mark

    Charlie,

    Thanks for the analysis.   Since the motivation for the proposed changes is to improve safety, do you have any estimate of the additional injuries/deaths caused by all those extra car trips?  

  • Jeff

    Two deaths over the course of a few weeks, and we’re considering adding considerable delays to billions of passenger-trips per year and spending billions of dollars on platform screen doors.

    Meanwhile, above ground…

  • Anonymous

    I think even adding platform screen doors would be cheaper than all the time lost by slowing down the trains, especially in the long run.

  • Larry Littlefield

    I wouldn’t assume the 12-9s are the actual issue.  For all we know it’s a 20/50 pension, though they can’t say that.

    Meanwhile, what does Fred Ferrer have to say?

  • Ridiculous. NYC has one of the slowest subway systems in the world as it is, with an average speed around 17 MPH, while many other systems average over 20 MPH. And slightly slower speeds isnt going to allow the driver to stop when a pax is pushed right in front of the train.

  • Average speed by system:

    http://www.flickr.com/photos/39017545@N02/4439291903/sizes/l/in/set-72157621837169293/

    NYC Subway and PATH are the two slowest systems as it is.

  • Joe R.

    Because trains will take longer to complete their runs, you’ll either need more trains for any given frequency of service, or you’ll have to decrease the frequency of service. The latter is what will happen because the MTA can’t afford to buy more trains and hire more train operators. So with less frequent, slower service you’re sure to lose some riders.

    And yes, the system is slow enough as it is, especially with the trains being “detuned” after the Williamsburg Bridge accident in 1995. Granted, some of the older lines, particularly the IRT, have many sharp curves, but most of the IND and BMT have no such issues. Locals should be able to average about 20 to 22 mph, and some of the better expresses should be able to average over 30 mph. This of course assumes the trains are ran to design specs. The newer trains can actually maintain their initial 2.5 mph/sec acceleration rate up to about 25 mph, and have a design speed of 55 mph (and I’ll bet with minor modifications they could go to 80 mph).

  • Anonymous

    @8f996ad67f04aec5edcfbc5070d76441:disqus The short answer is that my figures assume that NYC crash injuries (including fatalities) would rise in proportion with the predicted increase in overall NYC vehicle miles traveled, which is just 0.6%. So if there are, what, ~300 annual fatalities, that figure would rise by two. If there are 3,000 serious injury crashes (and there I’m really guessing), that would rise by around 20.

    FYI, my modeling of crash costs pivots off of the estimated baseline cost of NYC traffic crashes, which is around $4 billion a year. Add 0.6% to that and you get around $25 million, which you can sort of see in the chart: it’s the top band among the five bands comprising “Environment.”

  • Anonymous

    @8f996ad67f04aec5edcfbc5070d76441:disqus The short answer is that my figures assume that NYC crash injuries (including fatalities) would rise in proportion with the predicted increase in overall NYC vehicle miles traveled, which is just 0.6%. So if there are, what, ~300 annual fatalities, that figure would rise by two. If there are 3,000 serious injury crashes (and there I’m really guessing), that would rise by around 20.

    FYI, my modeling of crash costs pivots off of the estimated baseline cost of NYC traffic crashes, which is around $4 billion a year. Add 0.6% to that and you get around $25 million, which you can sort of see in the chart: it’s the top band among the five bands comprising “Environment.”

  • Anonymous

    @8f996ad67f04aec5edcfbc5070d76441:disqus The short answer is that my figures assume that NYC crash injuries (including fatalities) would rise in proportion with the predicted increase in overall NYC vehicle miles traveled, which is just 0.6%. So if there are, what, ~300 annual fatalities, that figure would rise by two. If there are 3,000 serious injury crashes (and there I’m really guessing), that would rise by around 20.

    FYI, my modeling of crash costs pivots off of the estimated baseline cost of NYC traffic crashes, which is around $4 billion a year. Add 0.6% to that and you get around $25 million, which you can sort of see in the chart: it’s the top band among the five bands comprising “Environment.”

  • Em

    I thought we all say attaching value to travel time is bogus, at least in discussing savings or loss of minutes in highway bypass and expansion projects?

  • Mark S.

    So when considering alternatives for reducing deaths, the platform doors sound a lot cheaper — a one time expense of $600 million (is that the number I read?) versus $1 billion in delays.

    Is it worth dealing with this problem? Society is generally willing to spend roughly $6 million to save a life. Fifty lives a year adds up to $300 million annually, and the value of avoiding injuries would be significant, too. This doesn’t match the $1 billion cost of slowing the trains, but platform doors look like a winner, with a one-time expense of less than $1 billion.

  • Ian Turner

    @0496b4f7904286e562a38f64d14f8a26:disqus : I would strenuously challenge the idea that society is willing to spend $6 million to save a life. If that were true, we would have universal healthcare, would eliminate fuel oil heating systems and constrain coal power, we would have bollards anywhere pedestrian areas and highways interact, and we would make it much more expensive to drive.

  • Ian Turner

    Oh, also, since the average (not median) person makes much less than six million dollars in a lifetime, it’s not clear how you would pay for that even if it were true.

  • Ian Turner

    @b1d12f2ee346329e36b26a93424dbef1:disqus : Who was saying that?

  • Anonymous

    The “value of a statistical life” used by federal agencies (such as the EPA and FDA) that conduct risk-benefit analyses is indeed around $6 or $7 million. See http://en.wikipedia.org/wiki/Value_of_life.

    Of course, this is never about saving a specific life, but only about estimated averages, and the estimates for lives saved by a given intervention can be controversial.

    This value doesn’t reflect the lifetime income of the average person, but what people in government believes society is willing to spend.

    (It goes without saying that whatever analysis a government agency conducts, the final decision may be political!)

  • Ian Turner

    @qrt145:disqus : It may be the case that certain government agencies *say* that the value of a life is around that point, but since there are many opportunities to save a life for much less social cost which are not taken, it is clear that the government’s actual value of a life must be much lower. Actions count more than words.

    Of course, it is not surprising that governments are irrational when it comes to the cash value of a life, since people are also (intensely) irrational with respect to this topic. Statistical lives are valued significantly less than specific individuals, for example. Nonetheless, there is no reason to think that the public transportation system should be subjected to a much higher threshold than the much lower implicit threshold which is used elsewhere in society.

    The relationship of the cash value of a life to average income is that society probably couldn’t realisticly afford to pay a price per life which is significantly higher than average income, if the cash value were used across the board.

    Finally, if we really are talking about the utilitarian allocation of resources, I should note that you can prevent a death due to malaria for just around $2,000, so setting the bar at $6 million implies that it is OK to kill about 3,000 developing-world children in order to save one American — a rather ghastly conclusion.

  • Mark S.

    Ian Turner: Regarding malaria: if it costs $2,000 to prevent a death, then malaria death prevention gets a pretty good bang for the buck, by any standard.

  • Joe R.

    @0496b4f7904286e562a38f64d14f8a26:disqus The cost to retrofit the entire system is way above $600 million. I think that $600 million figure may have been just to do a few key stations, perhaps even only ONE station. It would likely cost tens of billions to do the entire system. And it’s not a one-time expense. You’ll have ongoing maintenance to make sure the doors work. And this being NYC, if the doors have any glass in them, I’m sure that will get broken or scratched on a regular basis. You’ll probably also need to stop trains more precisely than now to line up with the doors. That will impact schedules somewhat (although not as much as entering stations at 10 mph). Add all these annual costs up and you’re probably looking at tens of millions.

    The only cost-effective idea I’ve heard so far to prevent track deaths are motion-detector systems which alert a TO if someone is on the tracks in the next station. You’ll need to make sure rats and passengers leaning over the platform don’t trigger the detectors, but that seems a lot easier than installing platform doors. Moreover, the detectors would be located in places where they’re not subject to vandalism.

  • Ryan Ng

    And this won’t waste years of commuters’ time because…?

  • Edward

    Why not set up barriers with sliding doors.    So when trains pull up, the driver will make sure it stop stop line so it can align up with the barriers’ sliding doors (the barriers’ sliding door will be wider so the train doesn’t have to center perfectly)    That way no one can be pushed into the track…   hmm? 

  • Joe R.

    @edcantrell:disqus The reason why not is the doors would cost tens of billions to install systemwide, and entail millions annually in maintenance. While saving lives is a laudable goal, there are limits to what society can and should spend to save a life. In this case we’re talking about a handful of deaths each year (I’m only counting deaths where the person is pushed, not incidents where someone intentionally goes on the tracks to retrieve belongings, or falls due to being inebriated). If measures to prevent some of these deaths drive up the fare, or slow down the trains, people will flee the system and drive or take buses. You’ll likely end up with more people dead overall, including cancer deaths from the additional motor vehicle exhaust. That’s the real issue here. You may prevent some subway deaths with draconian measures, but then you’ll have more people dying in other places as a consequence.

    Solutions which don’t cost much, and which don’t impact train schedules at all, are the only things we should be looking at. I feel the idea of motion sensors, if properly implemented, might meet both those criteria.

  • Ted King

    There’s a missing piece in this discussion – refuge niches. How many of New York’s stations have platforms that do NOT have a refuge niche under the edge of the platform ? The number should be ZERO but hearsay indicates that a sizable percentage of your platforms lack this basic safety feature.

    NB – I’m commenting from the San Francisco Bay area and am used to seeing refuge niches on both BART and SFMuni platforms.

ALSO ON STREETSBLOG

Streetsblog Q&A With TWU Local 100 President John Samuelsen

|
Last December, John Samuelsen was elected president of TWU Local 100, the union that represents 38,000 subway and bus workers in the New York City region. He assumed the leadership from former president Roger Toussaint at a troubled time for the transit system. With transit tax revenues in free fall and state lawmakers raiding MTA […]

Will Cuomo Help the MTA Make “Every Dollar Count”?

|
The MTA has made big strides in recent years to streamline its operations, but without political leadership from Governor Cuomo, the agency won’t be able to tackle the high costs and inefficiencies that continue to hamper the city’s transit system. In 2010, after an acute budget crisis, the MTA began a program to cut costs […]

Google Transit Mapping Comes to New York

|
This morning a number of public officials, including Governor David Paterson, joined Google co-founder Sergey Brin in announcing his company’s new city transit mapping tool. The service covers MTA subway and bus lines, along with PATH and New Jersey Transit routes. New Yorkers (including yours truly) who have been frustrated in the past by Google […]

MTA to Boost Frequency of Q70 to LaGuardia

|
The MTA announced this morning that it plans to increase service on the Q70 Limited bus to LaGuardia Airport and hopes to roll out Select Bus Service on the route later this year. Riders Alliance Executive Director John Raskin applauded the news as “a great step toward implementing a real airport shuttle from the subway” […]