Off-Peak Discounts for NYC Transit: An Intriguing Idea

Discounting off-peak transit service could be a boon to New York City’s transportation and quality of life, so long as revenues can be found to make up for the likely farebox shortfall.

MTA chief Jay Walder floated the idea of off-peak discounts in an interview in today’s New York Times. While Walder didn’t offer quantification, the Balanced Transportation Analyzer software model I’ve developed with Ted Kheel can estimate the effects of time-varied subway fares — not just how ridership might shift from peak to off-peak periods, but indirect impacts such as the shift of auto trips to transit and the resulting changes to car travel speeds.

The results look promising for this prototype fare structure that I tested with the BTA:

  • 1/3-off subway fare from 11:00 p.m. to 5:00 a.m.
  • 1/6-off subway fare from 5:00 to 7:00 a.m., 10:00 a.m. to 4:00 p.m., and 7:00 to 11:00 p.m.
  • 15 percent higher subway fare from 8:00 to 9:00 a.m. and 5:00 to 6:00 p.m. (Although Walder referred only to off-peak discounts, the model suggests that forestalling an increase in ridership during the two peak hours, when the system is strained beyond capacity, could require raising fares at those times.)
  • No fare change during the "shoulder" hours of 7:00 to 8:00 a.m., 9:00 to 10:00 a.m., 4:00 to 5:00 p.m., and 6:00 to 7:00 p.m.
  • 1/4-off subway fare at all hours on weekends and holidays.
  • 1/4-off bus fare at all times (not mentioned by Walder but assumed here to preserve overall fare parity).

Here are the results:

  • The average price of a subway ride drops by 23 percent, equivalent to a $210 annual savings for a typical straphanger who takes 12 trains a week.
  • Notwithstanding the overall discount, however, peak-hour subway users who could not change their commute times would pay $100 a year more in fares.
  • Annual savings of $230 for bus riders, due to the assumed 25 percent drop in bus fares.
  • Subway usage increases 3 percent, even as morning and evening peak hour ridership drops by 1 percent and 3 percent, respectively, slightly easing crowding during those critical times.
  • Bus usage increases 5 percent.
  • 15,000 fewer cars enter the Manhattan CBD on weekdays, raising average speeds there by 2 percent.
  • Car and truck drivers save six million hours of travel time worth an estimated $230 million that they now lose to gridlock each year — with a majority of the savings occurring outside the CBD.
  • A rise in cycle and pedestrian commuting due to lower traffic, with the resulting increase in physical activity translating into health and longevity benefits worth an additional $116 million a year.
  • Fewer crashes and less pollution, with health and related benefits close to $100 million a year.

The downside of this program is an estimated $300 million drop in farebox revenues: $134 million on the subways, $162 million on buses.

The logical place to make up the shortfall, congestion pricing, is a subject Walder will obviously want to avoid until he is on even firmer political footing. The synergies are strong from a technical standpoint, since differential subway pricing would help the subways absorb car drivers whom a cordon toll would induce to switch to transit. The political synergies could be strong as well if differential fares help expand the constituency for congestion pricing.

  • rlb

    The obvious question is how is this reconciled with unlimited metrocards?

  • JK

    Great to see the powerful BTA used to provide a rapid read on what these policy recommendations could actually mean. Doesn’t it seem that the most likely way these off-peak fares would be adopted would be during a fare hike? Then, the hike would consist of raising the peak fares and not the off-peak. If so, would the MTA then still raise more money than it currenrly gets? How much more? The big political problem with much of the BTA’s “cost savings” is that they consist of very diffuse reductions in externality costs, while actually costing the MTA real dollars.

  • TKO

    As a person who has to use the subway at peak hours I find the idea a poor one. Some folks have jobs that have hours that can not be changed or to be lucky enough to work from home.

  • Rootboy

    TKO – but some folks don’t have jobs like that, and they’re the ones who would get a big incentive to ride at different times, making your commute less crowded.

  • As a person who has to use the subway at peak hours I find the idea a poor one. Some folks have jobs that have hours that can not be changed or to be lucky enough to work from home.

    Yes, and those folks will have an easier commute, with less crowding and door-holding.

  • Shemp

    If this requires a new fare system like Oyster Card I won’t hold my breath.

  • lee schipper

    A good idea that came to London 20 years ago (cheap off peak rail and subway cards) and Washington DC as well. Of course some folks can’t change, but many can. To those who complain about paying more in cash at peak: would you rather pay by being squashed more? Would you not pay for a bit more room? Or should we all pay to enlarge capacity at peak just so we can have more empty trains off peak?
    Lee Schipper

  • by stander

    hmmmm. congestion pricing on the subway but not on the streets….

  • Erin

    I am highly skeptical. Mostly I don’t believe that there are so many drivers who would ride transit instead of driving just because the fare is 17% lower.

    As someone who has studies transportation engineering, I’ll say that you should be very suspicious of study results claiming a 2% rise in average car speed. Think of all the parameters and assumptions necessary to determine how many fewer cars would come to the city, how much the increased distance between cars would result in an increased average speed. Generally, transportation engineers’ tools (software) lend itself to producing the results you want to produce. Plus, ask yourself how much we want the vehicle speed to increase here. Isn’t 10 above the speed limit fast enough?!

    And, if you consider simplicity, this will be a logistical disaster. Any Metrocard daily/weekly/monthly passes would no longer function (unless it’s changed to a ‘pay after’ system where you’re sent a bill for the Metrocard use from the previous period).

  • Erin

    I meant “studied transportation engineering”, not “studies…”

    ‘by stander’ had a good comment above: congestion pricing on the subway but not on the streets? The priorities are all wrong. If reducing the number of vehicles on the streets and increasing bus use are purported goals of the convoluted transit pricing scheme, that would be better accomplished through vehicle congestion pricing, not transit congestion pricing!

  • Jason A

    Thanks to the shortsighted actions of the past generation, and all leaders at all levels of government, as is, the MTA is severely overburdened. The Lexington, Culver and Queens Blv. lines are much too overcrowded – they all run at +100% capacity during rush hours.

    Unfortunately, there’s not going to be any relief any time soon (I doubt we’ll even finish that half-assed version of the Second Avenue Subway…).

    Consdiering that, I think this is workable way to relieve congestion on the trains and speed up commutes. I really don’t see what the problem is, outside the depressing fact these are the (cheap) strategies we’re reduced to considering for expanding transit service in a city that demands so much more…

  • rlb

    “The obvious question is how is this reconciled with unlimited metrocards?”

    I’ll field that one.
    To have an oyster card style system where there are unlimited rides, each time an account is used the system would check to see how much money had been charged to the account in the past month. If the ride were occurring off peak, the total for the past month would have to be above a certain amount to allow for free travel. If it were during peak, that amount would have to be higher.
    This method would have advantages over the current metrocard for the user. The first is the initial down payment/commitment to a monthly pass goes away. Particularly good for the fairweather cyclist faced with the nearly month of continuous rainy days we had at the beginning of the summer. The system could even be tailored to grant unlimited for that particular day, weekend, week, or month – which ever comes first so to speak.
    Another advantage would be the continuity. The system would always check back 30 days from the use, so you can imagine swinging between unlimited territory and pay-as-you-go depending on your recent frequency of use.

  • If its anything like DC, it means fares would naturally go UP. Also, no system would use so many different prices, it would be peak and off peak, like metro north uses (and DC).

    DC also charges peak fares from midnight to 3am on friday and saturday (service usually ends at 12), which actually makes more sense then a discount.

  • Author comments:

    rlb: “how is this reconciled with unlimited metrocards?” I honestly don’t know. I’ll admit further that the BTA assumes per-ride pricing, and may therefore overstate somewhat the propensity to switch travel times from peak to off-peak. (Further confession: In running the model I non-conservatively inputted a 2% average “service improvement” to reflect presumed spreading of peak loads.)

    Erin 1: The model predicts that the fare structure I inputted would eliminate 15,000 of the current 870,000 daily vehicle trips into the CBD, or fewer than 2 percent. If that seems implausible, please look inside the model and report back which elasticity assumptions and algorithms you would change.

    Erin 2: Re the predicted 2% improvement in CBD travel speeds. The BTA algorithm for relating city streets’ traffic densities to speeds was suggested to us by the dean of U.S. transportation economists, U-C Irvine Prof. Ken Small. It’s presented in the worksheet, Speed-Vol., Cordon. Wouldn’t it be more helpful for you to take a look at it first and then criticize?

    Erin 3: Average CBD traffic speeds are around 7 mph during the weekday “twin peaks” of 6-9 a.m. and 2-8 p.m. (This is also in the BTA, in the Motor Vs worksheet, Row 145.) Do you honestly believe that raising this average would be bad?

    Erin + by stander: I’ve written 8 Streetsblog posts in the past year urging congestion pricing for vehicle entries into the CBD. This post was intentionally pegged to Jay Walder’s brave trial balloon for time-differentiated transit pricing. If you want CBD (or other) congestion pricing for motor vehicles, don’t waste your time lobbying me or this blog, lobby your elected officials.

  • Erin

    Article Author: You’re obviously taking the comments very personally. Mine were meant as general criticisms of transport prediction and analysis, not as specific criticisms about your methodology. Would it be helpful if I read the paper I didn’t know existed (and which I still am not sure how to locate, even after your terse response)? Sure, probably. Thanks.

    I don’t keep track of which articles you write, and I don’t really care whose are whose. I read articles on SB, period. I’m glad you’ve written several articles about congestion pricing for vehicles. I’m glad you’re writing about transit pricing options too. It just so happens that I don’t believe that a more complicated transit fare structure is optimal. Maybe we won’t end up with optimal… we probably won’t.

    We the readers have this handy-dandy comment box beckoning, and we use it. If the authors are going to make it a habit to be so aggressive if a reader doesn’t agree with everything, or is just making his/her questions and thoughts known (these are comments, not mandates!), then the perceived comfort with with readers comment is going to be compromised.

    Nobody’s “lobbying” you. That anyone was “lobbying” you in particular seems like a rather pretentious claim. I write to my elected officials too.

  • J:Lai

    anything more than 2 fares is probably too complicated (peak/off-peak) – this is an agency that can’t even put up system-wide signs to show next train arrival despite having successfully done it on the L train.

    RLB’s proposal for an unlimited fare based on some trailing 30-day average of use is both highly confusing and not at all desirable, if I am understanding it correctly. If I buy an unlimited pass, I want an unlimited pass, not a pass that might be unlimited depending on how many time I used it last week. And I definitely don’t want to have to solve algebra problems just to figure out if my pass will be unlimited today.

    If peak-pricing for transit is introduced, it will either run concurrentlly with unlimited cards, or it will require eliminating unlimited cards.

    In the first case, it only impacts people who don’t use unlimited cards. That would be mostly tourists, the poor, and infrequent riders.
    The first group probably has very little demand elasticity, as does the second. Infrequent riders are the ones who are most likely to change their commute times in response to price.

    In the second case, there is a lot more potentially to change behavior in response to price. However, giving up the unlimited cards would amount to a fairly large fare increase for regular commuters.
    If the typical monthly unlimited user pays around $1.70 per ride (give or take), then switching to full fare at 2.25 represents an increase of over 30%.

  • Christopher

    While I’m not fundamentally opposed to off-peak discounts (or peak premiums, depending on how you look at it), I don’t see how this will solve any problems. Do we have any numbers as to how many people take transit at peak times who have any flexibility as to travel time? If I had to guess, that number would be very, very small. Why? Because the disincentive is already built-in. The trains and buses are jam-packed at peak times and, while they may run somewhat more frequently, they don’t get you where you need to go any faster. I would think that anyone who can stay off public transit in NYC during peak hours already does so. They don’t need any financial incentive to do so. As for shifting users from cars to public transit, I doubt that the cost differential is going to be great enough to change anyone’s habits. If someone was going to drive into Manhattan during peak travel times, either they already have free parking (civil servants, arrgh), or they’re committed to a lengthy hunt for a meter, or they’re going to pay a high price for off-street parking. Saving 50 cents isn’t going to change their minds about taking the car.

  • Felix

    They could make it free overnight and it’s still going to suck standing forever on the platform after a night on the town. I’d rather pay double and have a shorter wait, though I’m not claiming the numbers would work.

  • egk7

    Re: Unlimited Metrocards. In Stuttgart Germany you can buy two sorts of passes: peak and off-peak. Off-peak passes don’t work for travel at peak, and so are used by, eg. students, retirees and others with flexible schedules (if they have to travel at peak, they buy a single fare). Usually you know if you are a peak or an off peak person.

  • metrocard

    Maybe I’m missing something, but if the cost of off-peak goes down, doesn’t that mean the cost of peak will likely go up to make up the difference?

  • gregor samsa

    “1/4-off subway fare at all hours on weekends and holidays.”

    I am sorry – have you been on the subway in the weekends recently? On the weekends, it is RUSH HOUR ALL DAY LONG. Crowded stations, crowded trains, slow service. This is effectively increasing ridership and and severely reducing revenue on what’s already a very stressed system. Not gonna happen.

    “The downside of this program is an estimated $300 million drop in farebox revenues: $134 million on the subways, $162 million on buses.”

    Oh, it’s always the little things innit? The fact that this nugget is not on the post headline (here’s a suggestion: “OFF-PEAK DISCOUNTS DROPS MTA REVENUES BY $300 MILLION”) but rather buried at the end of it shows how out of touch with reality the proposal is.

  • neil wilson

    Does anyone understand cost accounting?

    NYC pays a LOT for peak service. They need to pay for workers. They need to pay for subway cars. They need to pay for big stations and longer platforms. Etc.

    Off peak costs a lot less. Payroll is lower because you can’t just hire people for the morning and evening rush hour. You need to have them work other times too. The cost of the cars is almost $0 because you bought the cars for peak times. The cost of the stations are close to $0 because they too were built for peak times.

    So, the marginal cost of off peak is far lower than what anyone is considering. If it were me, I would have fares at $1 off peak and have them slide up to $3 during peak times.

    I wish someone would put together a reasonable cost analysis of what it costs to run a subway during peak and off peak times.

  • Joe

    Sorry but this is a bad, BAD idea. The only people who use pay per ride Metrocards are tourists or those who don’t use the system often enough to warrant an unlimited card. Those who can’t afford the unlimited passes seem to be a very small minority. We need the income that comes from pay-per-rides and it can’t be replaced from elsewhere. And it shouldn’t be when the benefits are not there. We shouldn’t encourage more people to ride the subway late at night. Service will not increase so there’s not enough room for more people. Late-night trains in some areas are already packed. Also we don’t want to increase capacity at night or on the weekend because this is the time when real repair work is done!

  • Joe, I am a NYC resident using a pay-per-ride metrocard about 5-10 times a month because most of the time I’m bicycling but sometimes I bring my bike on the train or I don’t bike.

    Personally, the unpleasantness of the subway during rush hour is one of the main factors keeping me on a bike. Differential fares would give those with flexibility the opportunity and incentive to shift to an earlier or later commute. But parents have very little flexibility as to when they get their kids to school; often it is a 15 minute window. I would be reluctant to hit parents with an increased fare for riding during 7:30-8:30. But distinguishing parents with kids from others is just another wrinkle that I’m not sure the technology can handle.

    Frankly, I think parents escorted by a child should be permitted to use their kid’s free MetroCards. Making the parent pay to take the kids to school is just one more factor pushing the parent to drive instead of using mass transit.


Kheel Plan 2 Seeks to Plug MTA Budget Gap

Ted Kheel and his band of transportation analysts are releasing an updated version of their low-cost transit proposal, which they are pitching as an alternative to the Ravitch Commission’s MTA rescue package. The revised Kheel Plan retains the original’s congestion zone cordon, charging vehicles to drive into Manhattan below 60th Street. The major twist is […]

This Week: Street Safety, Transit Experts, and a Trip to Albany

It’s a big week to get involved with neighborhood efforts to improve street safety: A community board in Bedford Stuyvesant will mull a 20 mph neighborhood speed zone, Brownsville residents will figure out where they want better walking and biking, and the Boerum Hill Association will host a traffic safety forum, followed by a Park Slope […]

This Week: MetroCard’s Future, Jamaica Bay Greenway, Move NY

It’s a busy week for livable streets. Tuesday features a bevy of community meetings in Brooklyn, Wednesday continues with discussions of play streets, MTA fare payment technology, and Vision Zero in Astoria. Stand up for better bike and pedestrian routes in Manhattan and Queens on Thursday, and cap off the week with a discussion of […]

This Week: 14th Street PeopleWay, Remembering Asif Rahman

A very busy week on the Streetsblog calendar starts off with a memorial ride later today for Asif Rahman, who was killed by a truck driver while biking on Queens Boulevard eight years ago and would have turned 31 today. Queens Boulevard is getting safer thanks in large part to the advocacy of Asif’s mother […]

This Week: Traffic Calming for Co-Op City and Park Slope

This week’s calendar — particularly Tuesday evening — is packed with livable streets events, including community board meetings for traffic calming in Co-Op City and Park Slope. Here are the details: Monday: Join Park Slope Neighbors for a happy hour to benefit Riders Alliance and its campaign for better bus and subway service across the […]