that second on-board employee

A reader asks:
My question relates to the relationship between frequency and capacity.  In Boston on the MBTA … for many of the trains, there are 2 employees running the train.  On the Green line, trains are 2 cars long, with a driver in the first, and in the second an operator responsible for opening and closing doors and making sure no one gets on without paying.  For the other lines, the 2nd operator only has to open and close doors because you need to pay to get into the stations.
To me, it seems like a waste to have to pay a second individual to open and close the doors.  Outside of the highest frequency travel times, and even possibly during those, wouldn't it be better for travelers to have service twice as often even at half the capacity?  Outside of the truly busy travel times, trains rarely run anywhere near capacity.  Especially on weekends and in the evening, trains are never full but the less frequent service does not encourage those spontaneous transit trips that are so vital to urban life.
This is not my core expertise, but my understanding is that generally this is right:  the second employee is usually a holdover from days when fare collection and monitoring of doors had to be done manually.  The job often survives because it's coded into labor contracts and sometimes also into regulations.

I am unaware of anything that non-driving on-board employees do that would be utterly impractical to automate today, the best evidence for which is that trams, streetcars, light rail, and heavy rail can be found operating with a single employee all over the world.   (Fully grade separated heavy rail, of course, can also be run with zero on-board employees, liberating the agency to operate intense frequency even late in the evening.)  Fare collection is increasingly handled by Proof of Payment systems which feature roving fare inspectors.  While these fare inspectors have a cost, their number is not directly related to the number of vehicles in service, so they are not such a direct barrier to increases in service. 

Frequency is driven by staffing requirements rather than vehicles, so the number of employees on board is the dominant variable determining how frequently any line can be run.  Only during the peak commute period is the availability of vehicles a significant element of the frequency decision.
As you would expect, however, any local debate about turning second employees into drivers of additional service will be fraught.  It is very easy for opponents (usually including the unions) to make generalized allegations about safety and security because most people feel safer and more secure if there's an employee nearby.  So it's politically hard to do.

This is one of those issues that is intensely local, and where examples of experience from other cities just have trouble penetrating a local debate.  It happens even in Europe.  See for example the peculiar fare-collector job that exists on Amsterdam trams.  A little cubicle placed at the middle of each tram contains an employee who serves as a cashier, selling tickets.  Boarding and circulation on Amsterdam trams is awkward, and effective capacity much reduced, because you're required to board only at certain doors and exit at certain others.

This second employee on Amsterdam trams is, as near as I can tell, unique in Europe; everywhere else trams run with one employee (the driver) and roving fare inspection.  Get a European transit professional going on how bizarre this Amsterdam practice is.  It's great fun over a beer.  But they can also explain, politically, why it will probably never change. 

If readers know of recent stories where second employees have been successfully removed and retrained as drivers, thus allowing more service, please post a link in the comments.

44 Responses to that second on-board employee

  1. Eric Fischer February 15, 2011 at 9:50 pm #

    I don’t know how recent examples have to be to count as recent, but it was in the 1990s that the CTA rapid transit lines in Chicago were converted from two-man to largely one-man operation. Conductors continued to be employed in the subway sections, boarding at one end of the subway and disembarking at the other to wait for the next train in the other direction.
    The conversion involved roping off the former “railfan seat” at the front of the first car of each train so that at center-platform stops, the motorman could leave the cab and go over to the other side of the train to check when it was safe to close the doors.

  2. EngineerScotty February 15, 2011 at 10:25 pm #

    Portland’s WES commuter rail uses two-man crews due to FRA regulations.

  3. francis February 15, 2011 at 10:39 pm #

    a thought…
    – the fare collector person probably gets paid less than the driver. So if having one speeds up boarding and improves service and the fare collection rate enough, the cost might not be so great.

  4. Nicholas Barnard February 15, 2011 at 10:55 pm #

    Its not quite the same business, but airlines successfully went from three person cockpits (Captain, Co-Pilot, and Engineer) to just two person cockpits (Captain, Co-Pilot). This took a little while to fully accomplish (the 737 was the first notable jet with only two people in the cockpit, and the 727 was probably the last passenger jet out of the air, although the DC-10s made it even further as freighters.
    Most of this came as a result of lots of automation within the cockpit.
    Of course the question also arises do they have enough vehicles to increase frequencies?

  5. ant6n February 15, 2011 at 11:15 pm #

    Berlin’s S-Bahn successfully went from two people per train (with 8 cars) and all-staffed stations to just the driver, and staffing only on the most busy stations.

  6. In Brisbane February 15, 2011 at 11:49 pm #

    Brisbanes trains have 2 people, a driver and a guard.
    The platforms are uneven/have different heights and so there is a gap (both horizontally and vertically) and so they need to be there when putting out ramps for the disabled etc. Not sure if this can be automated though.

  7. Daz February 16, 2011 at 1:48 am #

    Melbourne’s trams went from driver and conductor to driver and ticket machines in the early-mid nineties. The Liberal party Kennett government came to power with plans to privatize just about everything, including the trams and they were largely successful. Dropping the conductors was a prelude to privatization. In the end it was an easier task than they expected. From memory, they had plans (and forklifts) ready to deal with tram drivers parking across evey intersection in the cbd, but it all happened with a whimper rather than a bang.
    Brisbane trains still cleave to the old style, plate the train guard must signal it’s safe for the train to leave the platform. I believe the guard also controls the doors, and on unstaffed platforms they deploy the whellchair ramp. Same for Sydney, I believe. Melbourne, I couldn’t say. I’m always too distracted by their immodest 5 foot two inch gauge to notice staffing arrangements.

  8. Tom West February 16, 2011 at 6:40 am #

    London Underground services used to have a guard until newer rolling stock allowed the drivers to do the doors. Some of the suburban heavy rail services also have diver-only operation, but these tend to be frequent service with many stops, like a subway.
    On other hand, most other rail servies in the UK have a second staff member. These services tend to be at least of long distance, infrequent service or infrequent stops. Therefore, the penalty for getting off at the wrong stop is much higher compared withj the subway-like suburban services around London.
    I think the main argument in favour of having someone other than the driver on a rail service is for customer service, and this matters most where mistakes by passengers have the biggest negative impact. (Going 10 miles past your stop and having to wait an hour or more to go back the other way…)

  9. Jeff Craig February 16, 2011 at 7:12 am #

    Your anti labor posts are getting more often. I thought u were pro transit. Guess I was wrong.

  10. ant6n February 16, 2011 at 7:59 am #

    I think this is about trying to get the most service out of a given set of employees. Transit is not a work stimulus program.

  11. anonymouse February 16, 2011 at 9:39 am #

    Peak frequency is often limited by available vehicles or line capacity, but off-peak, frequency is indeed “driven by staffing requirements rather than vehicles”. And it was my understanding that in Boston, the conductors on the Green Line were not just there to collect fares but also to keep people from trying to walk through the gap between the cars, which can also be solved with mechanical barriers or just declaring that if they’re stupid enough to try, they deserve what they get as a result.

  12. Chip Olson February 16, 2011 at 10:14 am #

    Er, oops. Let’s try that again, shall we?
    Boston’s Orange Line went to single-person operation in the past year. They installed mirrors and monitors at the end of every platform to help the driver see the entire train. This was not accompanied by an increase in frequency, however; it was done as a cost-saving measure.
    I’m confused by anonymouse’s comment, as it’s impossible to walk between cars on the Green Line; they’re light rail vehicles with no end doors. The Green Line runs in subway downtown, with faregated stations, but at surface stops, if you board at the back door you’re expected to come up front to pay; I don’t know if anyone actually does. Some stations have people on the platform at peak hours to do fare validation. A card reader at the back door plus roving fare inspectors would make the whole business much more efficient.

  13. EngineerScotty February 16, 2011 at 10:17 am #

    The question of how to balance the needs of transit’s riders and needs of its workers, and of taxpayers in general, is a complicated one. In areas where revenue is fixed (most North American transit agencies lack plenary taxing authority, for instance), the relationship between labor and riders can quickly turn adversarial.

  14. ant6n February 16, 2011 at 10:27 am #

    At which point it might make sense to privatize operations and create open bidding processes for the network. Whoever offers the highest quality service given the available subsidies, gets the contract to operate the network (or parts of it). Add in some penalty clauses if the quality dips, and all of the sudden the transit agency is much less ‘socialist’.

  15. Transitmatters February 16, 2011 at 10:46 am #

    Two of the other lines operated by the MBTA in Boston, the Orange and Blue lines, have gone to single person train operation. Many have raised concerns, especially over the Orange Line going SPTO, with regard to budget priorities and the usual fear of increased criminal activity.
    The specific example of the Green Line is different from the heavy rail transit because 4 lines get condensed into two tracks that run through downtown. Higher load times due to high passenger volumes make it difficult to run the trains more frequently through downtown. Only the Southbound Park Street and both Kenmore platforms enable simultaneous loading of two trains on two separate tracks, but the bottleneck always ends up being the Northbound Park Street platform because of heavy foot traffic and since trains must load sequentially.
    The optimal solution would be for the MBTA to order longer, conventional light rail trains at the same length as 3 car trains being piloted on two of its lines (which has been discussed ad infinitum on the forums), but because of Boston’s segregated residential and commercial zones due to the zoning laws created mid-20th century based on false perceptions of improved quality of life, ridership falls sharply outside of peak hours and the MBTA takes this as an opportunity to exercise the originally intended flexibility of this format of light rail vehicle (much to the chagrin of many Bostonians who find themselves on crowded poorly-timed off-peak single-car trains).
    Alas, the MBTA has many other budgetary concerns and improved throughput on the Green Line would be better found through improved infrastructure and rollingstock that is less likely to fail and cause delays. As Jarrett noted, these sorts of things are moreso restricted by staffing requirements than availability of equipment and that’s the case we have here in Boston. Higher frequency at the same capacity would be ideal, but if you’re looking for cost savings through lesser capacity but higher frequency, you’d still have to contend with higher load times as more people try to cram themselves into the train in front of them rather than wait for the train ‘immediately behind’.

  16. D February 16, 2011 at 10:56 am #

    What about the energy/fuel and extra wear/tear on rolling stock that would result from to “hav(ing) service twice as often even at half the capacity”? If the labour costs can be slashed by half (2 employees down to one), are the energy/fuel and maintenance costs relatively negligeable; eg, not a barrier to doubling frequency?
    Not trying to be discouraging, but merely curious. I’ve often thought that with some commuter train lines (in Montréal’s West Island, or in the Fraser Valley in Vancouver), it’d likely be far better to have double the frequency with half-size trains… what combination of staffing, energy/ fuel/ maintenance, or running rights on freight lines makes this challenging?

  17. Leigh Holcombe February 16, 2011 at 11:42 am #

    In Quito, Ecuador, some of the bus drivers had “assistants” whose job was to collect fares and keep an eye out for potential riders. The assistant seemed to be totally informal, taking a cut of the fares in exchange for his services. Based on my limited experience in less developed countries, it seems like any job that could theoretically be accomplished by an additional worker will turn into an under-the-table opportunity for someone. Considering the amount of work the driver had to do just to navigate the chaos of downtown traffic, I’m glad the assistant was there.

  18. Nicholas Barnard February 16, 2011 at 11:52 am #

    @ant6n in theory that makes sense, but you get severely underbid contracts where the service provider can’t actually deliver the service. In theory its nice to simply rebid the contract if they can’t deliver, but in practice its a huge pain.
    A good example of this is the Major/Regional airline relationship. Take a look at Mesa. They’re consistently a low bidder, and most of the majors are doing their best to rid themselves of Mesa’s services..

  19. ReeD February 16, 2011 at 12:11 pm #

    I guess I’m confused. While the article focuses mainly on the relevancy of the second driver, it seems the more interesting point here is doubling the frequency. If you switch from 2 car trains every n minutes to 1 car trains every n/2 minutes, you have the same capacity but with twice the frequency. This doesn’t advocate any adjustment in number of vehicles used or in the safety / fare collection policies (doesn’t the MBTA already run 1 car trains with just a driver?)

  20. TransitPlannerMunich February 16, 2011 at 12:35 pm #

    Amsterdam re-introduced the conductor box several years ago, cause they thought they cannot get their passengers under control. A similar thing they have at their metro stations. Originally a proof of payment system they have now several ticket inspectors standing at every station. So far as I know people in some social benefits/welfare/unemployed program who work their for very little money – and also look like that. Some from my impression look like typical panhandlers with a plastic uniform.

  21. Felix the Cassowary February 16, 2011 at 1:37 pm #

    @Daz, Melbourne’s trains have just a driver. This is why some of the rolling stock is confined to just the ex-Hillside part of the network: In the ex-Bayside part, the screens that let the driver see whether its safe to close the door, are positioned so they can only be seen by someone sitting in the left side of the cab. But with the X’Traps, the driver sits in the centre of the cab.
    I’m told one of the reasons trams lasted in Melbourne longer than in other parts of the country is because the tram/bus drivers union insisted that buses should have a conductor, so there weren’t any staffing advantages to switching services. But buses have been single-staffed for longer than trams, so I don’t know how true this is.
    It seems absurd to me that anyone should want to run a train with two people, if one can do the job just fine. The point of public transport is to get people places, not to give unions want they want.

  22. Pete February 16, 2011 at 1:44 pm #

    London Underground is totally OPO (one person operated), with two lines currently having ATO (automatic train operation), these being the Victoria Line from opening in 1967, and recently the Central Line since 1999? (the 1992 stock trains being delivered with ATO capability). The Jubilee Line is currently being converted to ATO. In all three cases a driver is present to operate the doors and take over if the ATO system fails.
    The Docklands Light Railway (DLR) is also ATO but here the member of staff circulates in the cars checking tickets, and operating the doors.
    Regarding Amsterdam, conductors were reintroduced as TransitPlannerMunich rightly says due to endemic fare evasion on the open boarding system. The conductor’s booth is adjacent to the 4th doorway on the Combino cars.
    Whereas on the older cars the conductor sits by the rear doorway.
    I think there are advantages to having another pair of eyes to supervise passengers. I have witnessed a conductor turning away a passenger trying to board carrying a hot drink, and also telling a passenger talking loudly on a mobile phone to be quiet!

  23. Pete February 16, 2011 at 1:48 pm #

    The Tyne and Wear Metro serving the cities of Newcastle, Gateshead, and Sunderland in North East England is also ATO. This was the UK’s first use of light rail technology, with the cars being based on the tried and tested German Stadtbahn B design.

  24. Pete February 16, 2011 at 2:04 pm #

    Correction, the Tyne and Wear Metro is OPO NOT ATO.
    Apologies for the typo!

  25. anonymouse February 16, 2011 at 2:48 pm #

    @Chip, I meant crossing the tracks behind the first car of the train and in front of the second car of the train by stepping on or over the couplers connecting the two cars. It ought to be a very self-evidently stupid thing to do, but I’ve heard it used as the justification for the crew arrangement on the Green Line.

  26. anonymouse February 16, 2011 at 3:00 pm #

    @ant6n what you propose would probably make sense, but that’s not how privatization works around here. For example, Caltrain is currently looking at bids for its operating contract, but the contract specifies the timetable, fares, and even the staffing levels. The bidders have almost no freedom to actually make things more efficient and save any money other than by paying each employee less. And I would much rather have one well-paid conductor than two poorly-paid ones on my train.
    As a side note, it takes three people to run a Caltrain. One conductor is there to open and close the doors and not check tickets, while the other one is there to keep the first conductor from getting bored or helping customers by engaging him in conversation for the length of the trip.

  27. David Oleesky February 16, 2011 at 3:09 pm #

    Guards/conductors are needed where all 3 of the following conditions are met:
    a) stations are not manned or if manned are “open access” without ticket barriers;
    b) high capacity vehicles (e.g. trains, trams or double-deck buses) are used and “pay as you enter” would cause too many delays; and
    c) automated ticket machines at stops are likely to be vandalised and the money they contain stolen.
    This applies on much of the UK’s rail network and some tram systems (e.g. Sheffield).

  28. Daniel February 16, 2011 at 6:44 pm #

    Re: Melbourne, it wasn’t a direct link, but the removal of conductors from trams and guards from trains in the mid-90s was followed in the late-90s by a virtual doubling of Sunday services (at least, between about 11am and 7pm). But it’s not clear if the same staff were re-deployed, or new drivers hired.

  29. Zoltán February 16, 2011 at 7:14 pm #

    Helsinki’s suburban trains have a fixed staff level of two people, a driver and a conductor, be they 2, 4, 6 or 8 cars long.
    Despite this, they still like to maintain their ten-minute frequency on most routes in the off-peak by reducing trains to 2 cars long. That’s more expensive because of staffing, but I guess they feel it justified in terms of more people taking the train if it’s more frequent.
    Meanwhile, in Baltimore, their light rail trains have a fixed staff level of one driver, and run every 15 minutes (every 30 on each of two branches) with two articulated cars in the off-peak. I consider every 15 minutes a quite pitiful frequency for a light rail line running downtown, and it renders it quite useless for travel within downtown. The same strategy, of running single cars off-peak, could improve it to a respectable 7/8 minutes.
    The difference, I feel, between Helsinki and Baltimore is this: Helsinki is willing to spend money to make sure transit is attractive. Baltimore seems to view transit as meeting the social service needs a fixed minority of people, who are mainly poor and of a minority background, as cheaply as possible.

  30. ant6n February 16, 2011 at 8:26 pm #

    The contract requirements should include schedules and capacities. Don’t know about staffing. This is how it’s done in Germany, and it seems to work well. Just recently, DB Regio asked for 125 Million Euro to keep a regional train network running, 67% than before. So the transit agency opened up a bidding process, with some relaxed requirements to get some smaller competitors to bid. One such company, Trans Regio, asked for 88 million, and then DB Regio got the contract for 80 million.
    So opening up the process to competitive bidding is about getting smarter/cheaper operations, for example by removing unnecessary staff — but it is also, and even more importantly so, about paying actual market value for the services.
    Which, btw, doesn’t work if
    a) there’s really only one possible bidder. For example if only one bidder has the required rolling stock (in which case maybe the transit agency should own it and give it to whoever gets the contract).
    b) there are no effective penalty clauses, allowing the operator to let the quality slip. In the States there seems to be an issue with the cheapest bidder turning out to be the most expensive – that shouldn’t happen.

  31. Thom February 16, 2011 at 9:22 pm #

    The other night, an MBTA inspector told me that the MBTA is working on going to single-driver operation on the Red Line, too
    I’ve always wondered whether a roving fare inspector system could actually work on the Boston Green Lines. This is the busiest light-rail system in the US and often overcrowded to the point where I can imagine anyone “roving”. Certainly, they couldn’t get through the whole car in between stops. Yet, I’ve seen the fare inspector system work quite well in plenty of other places.
    Anyone know if there a rule of thumb that roving fare inspection gets ineffective beyond a certain level of utilization?

  32. Jarrett at February 16, 2011 at 9:24 pm #

    @Thom.  There are many rules of thumb, but understandably, agencies don't like to announce their inspection rates.  It undermines the deterrent effect. 

  33. francis February 16, 2011 at 10:09 pm #

    the businessperson in me thinks the following:
    – it makes sense to add a person (or maybe even multiple persons!) so long as they bring in more revenue than it costs to employ them. If fare is $2 and salary and benefits is $20 an hour, it’d take an increase of just 10 riders and hour, or 80 a day, for the added person to be profitable. (The Ecuadorean example comes to mind).

  34. Alon Levy February 16, 2011 at 10:23 pm #

    @D: it’s normal to run shorter trains in the off-peak. It’s doable if it’s possible to run trains with only one employee; it’s a much bigger problem if, like on the LIRR, each train has 5-6 employees.
    @Jarrett: Tokyo Metro and Toei went OPTO surprisingly recently – they started eliminating the conductor line by line by introducing new signaling in the last two decades. Toei is a government agency, as was Tokyo Metro until 2004, so it’s not just about privatization making it easier to eliminate employees.

  35. Jack Horner February 17, 2011 at 1:45 am #

    Melbourne, Adelaide and Perth trains are one person, not sure since when, I imagine within the last 20 years.
    I guess that the variable per train kilometre cost of running a train in Sydney is about double what it is in Perth (twice as many crew, trains about twice as big and heavy, therefore twice the energy and variable maintenance costs).
    Sydney has many line sections with a normal half hourly off peak service which could fruitfully be increased, if they could get their act together to run an efficient operation on the cost side.

  36. ant6n February 17, 2011 at 9:54 am #

    I’ve seen fare inspectors take positions at all exists of a subway station, and ask for proof of payment as people leave the station. That can be done when trains are really full, but might not work so well on surface stops.

  37. Jarrett February 17, 2011 at 1:31 pm #

    @ everyone talking about Ecuador.
    This post is about staffing calculations in the high-wage developed world.
    In countries with cheap labor the calculation is completely different. In fact, in India I recommended against automated fare collection, since there’s such an abundance of inexpensive labor to do these jobs.

  38. Bob Davis February 17, 2011 at 10:05 pm #

    Brings to mind the situation in San Francisco back in the 1940’s. When Muni bought out Market St. Ry. in 1944, it wasn’t long before nearly all the MSRy streetcar lines were converted to bus service. Ever since the 1930’s, SF had a city ordinance requiring two-man crews on streetcars, but not on buses. When this was finally repealed in the 1950’s, there were very few rail lines left, and most of the ran through tunnels where buses were impractical. One should be aware that SF is a strong “union town”, so even though logic would tell us that driving a bus is probably more challenging than running a streetcar, it took a long time for this to sink in. About the only remotely good thing to come from this was the electrification of many of the former streetcar lines for electric bus service. Since the City has its own hydro-electric plant, this makes the “fuel” for the trolley buses quite inexpensive.

  39. alurin February 18, 2011 at 10:41 am #

    ant6n: The MBTA commuter rail service is run by an independent contractor on a competitive basis. The MBTA owns the rolling stock. Still, the system doesn’t really work very well. The T tends to stick with the same operator despite inferior service, presumably due to the costs of switching operators. Recently the Globe published a story about the lack of reliability on the commuter rail. It turned out that it was in the contract for the contractor to replace rails and ties, but not to properly maintain the engines. Thus, there were few problems with track, but lots of problems with maintenance. The private contractor isn’t going to do anything they’re not required to do, and that has to be specified in detail; you can’t just say “Provide good service”. When it turns out you’ve forgotten to put something into the contract, you have to wait until the next contract comes up before renegotiating it. It’s just not an efficient way to manage a transit service.

  40. david vartanoff February 21, 2011 at 11:34 am #

    It is NOT anti union to question a full pay job which involves less than full work. In SF the second and third (when Muni still ran 3 car trains) drivers were often to be seen reading. While there might have been a use on the surface segments before POP was instituted, they were NEVER necessary in the fully barrier fare controlled high platform subway.

  41. Karl June 30, 2011 at 1:16 pm #

    There are two because in older green like cars you cannot open the second car’s doors from the first.

  42. Jeremy September 4, 2011 at 7:06 pm #

    Just for clarification, the MBTA reduced orange line frequency during the off-peak when they went to single operators. But in the peak, since they don’t have any more cars to use, they do “drop-backs” in which an operator is waiting in position on the platform at the terminus so that he/she can jump in and go (rather than have the existing operator walk all the way down to the other end).
    The green line has fare gates in the subway stations (Kenmore and east) but does on-board fare collection (to the driver) on the surface. This causes tremendous delays and needs to end asap. It seems to me that a proof-of-payment system could convert that second operator to either operate another train set or be a roving fare inspector. But in the peak there are no more cars available. Right now, the operators alternate driving in each direction, although still only one at a time.

  43. Jeremy September 4, 2011 at 7:09 pm #

    And when I say reduced frequency, I mean increased service, as in “every 10 min to every 8 min”. They did not double it, for two reasons. Some stations are curved/congested so they use “platform monitors” who signal the operator with a flashlight for door closings. Also, they want to save money; lots of people are retiring so they do this through attrition.

  44. greg byshenk December 27, 2011 at 10:05 am #

    I know that this is coming in very late, but I’d point out that there are conductors on only some Amsterdam trams. I believe they are on the lines frequented by tourists (who often need additional direction) and on lines where passenger behaviour is a problem. I don’t recall if it is still the case, but conductors were also added to some lines in Rotterdam a few years back.