How often have you heard that rush hour is when transit really excels? When you see all those crowded buses and trains, for just two hours or so, it seems like transit’s really proving its value.
Not so fast. Transit systems that run at a low level all day but then ramp up hugely during the peak can be very, very inefficient. That’s because putting out a bus or train to run just for 2-3 hours entails the following big marginal costs:
- You must get someone to report to work for a short shift. This usually means that you pay them for more hours than they work. Some agencies are also prohibited, in their labor contracts, from using part time drivers, which are really the only way to serve these short shifts efficiently.
- Peak demand that flows in only one direction, as in the classic American single-centred city, also generates the huge inefficiency of moving all those vehicles, entirely or mostly empty, back in the reverse-peak direction.
- The peak determines the number of vehicles that must be owned and maintained.
Sometimes, peak ridership is so much higher than midday that the fare revenue makes up for all these inefficiencies, but not always.
In fact, my experience with American bus operators is that few of them have really counted the cost of their peak-only services. Transit agencies should know and report the true marginal cost of a peak-only vehicle hour as distinct from an all day vehicle hour (including all-day buses that continue through the peak.) We know peak-only service is expensive, but we’d have much clearer conversations if we knew exactly how much.
For a real world example of such a study, see the Toronto Transit Commission’s’ express bus study
(https://www.ttc.ca/About_the_TTC/Commission_reports_and_information/Commission_meetings/2017/June_15/Reports/6_Express_Bus_Network_Study_combined.pdf)
In summary, peak express bus service have a median cost of boarding of 10.22, vs all day express median of 2.49 to 2.55. The only saving grace for those peak buses is that the subway in Toronto is beyond capacity and can not absorb those passengers on the peak buses.
Yet another factor that makes peak-hour service more expensive – peak hour service is when traffic is at its worst, so the same trips pay longer and require more driver-hours. On top of that, the amount that traffic lengthens a trip is highly unpredictable, so reliable service require lots of padding. Which means you’re effectively paying drivers for the worst-case traffic conditions every single day, even on days when traffic isn’t actually all that bad.
The Seattle area where I live has seen a large increase in revenue for the transit agencies over the past few years due to a booming economy. Unfortunately, a booming economy also means worsening traffic. So, much of that new revenue gets squandered adding more and more schedule padding to the existing peak-hour express runs, rather than actually adding more service.
In the late 70’s Edmonton Transit adopted a three-part costing formula which produced reliable results: (hours x hourly rate) + (kilometers x km rate) + (peak buses x annual cost per peak bus) = total cost per route. Winnipeg reportedly got even better results by calculating immediate costs and long term costs separately.
The push to do this came from the oil boom. The system’s peak-to-base ratio grew rapidly enough that the problem could be seen in all departments. (I used to board a bitterly cold trolley coach in the winter AM peaks, because they could be parked outdoors and still start in the morning. There was no immediate cost of garaging the growing number of buses, but eventually there was the long-term cost of building a new garage.)
This method was replaced with the old all-inclusive hourly rate when the energy crash occurred, because the more accurate approach did not fit the theory held by a new administration that off-peak service was a gold mine that could be used to protect white-collar peak commuters from cuts.
I believe that most systems using a single hourly rate do so out of tradition. Streetcar companies copied railway pen and paper accounting, often using costs per mile. Seattle Transit did so into the 1960’s. When a system only undergoes minor changes it’s hard to see anything wrong with single rate approaches.
We use a similar three part formula to compensate our bus operating companies in the Helsinki region: hours, km and daily peak vehicles. Bids to operate are submitted accordingly.
For tram and metro the formula is otherwise similar, but vehicles are compensated annually based on the actual rolling stock owned (as agreed) and the daily rate is merely a statistic. This presents some problems when calculating the marginal cost of service.
So a new peak service will usually incur a full vehicle day cost. Doesn’t properly account for a higher wage cost, but works for a ballpark estimate.
Isn’t this the reason why LADOT Commuter Express was created in the 1980s, taking over peak-only routes previously operated by RTD (the predecessor to Metro)? Commuter Express routes have their own fleet, separate driver contracts, and zone-based fares.
One anomaly however is Commuter Express 142, formerly operated by Long Beach Transit: it has daily, all-day, half hourly service, with just one fare zone and an open-door policy (passengers can get on and off at any stop). Route 142 thus operates like a regular transit route, and the only thing that sets it aside is a long stretch across the ports of Los Angeles and Long Beach without any stops. I never understood why it is operated by Commuter Express rather than Metro or Long Beach Transit.
The driver issue can be relieved a bit when the system is not too small, by smart assigning of services. A consequence is a split shift (morning a few hours, afternoon a few hours, but because that would be one or two single days in a sequence, it can level out.
(note: a sequence is five to six days beginning with a late shift, going through mid-day and split, and ending with a early shift; then there are resting days, and a sequence begins again)
You can sometimes do this, but split shifts often come with penalties in collective bargaining agreements where the time in between has to be paid at some rate, or a variety of other arrangements that all result in paying more for these split shifts than you would for an equivalent number of straight hours. Not that this can’t be done, it just means the peak hours cost more, as Jared said.
It can also be harder to hire drivers for split work, depending on the labor market in your area. Again, certainly can be overcome, but it usually means a higher pay scale compared to an equivalent position with nicer hours.
Back in the 1970s I interviewed the general manager of Calgary Transit about their upcoming C-Train LRT line. The one comment he made that I will never forget was that 10 years ago it took a bus just over 30 minutes to travel the McLeod Trail in the rush hour while it takes over a hour in the 70s. He said; “We are not buying buses to transport passenger but to store them.” i always thought that was an apt description of the effect of traffic on transit service.
Can you look at it the other way? That is, that if you are having to provide that type of expensive peak-only service, it’s an opportunity to boost your midday service at a lower incremental cost? After all, you already have the buses, you don’t have to schedule premium split shifts, and you don’t have the lost deadheading time to travel to and from the garage after the AM rush and before the PM rush.
In cities across the country, there are lots and lots of routes that run more often at noon on a weekday than noon on a Saturday, with similar ridership. I’d imagine that this way of thinking is a big reason why.
Depends what efficiency we are talking about. Lately you have been talking about geometric space efficiency, and it is during peak times that buses then really are space-efficient. So it depends on the problem we are trying to solve.
There are some values to the peak services to nom-users. This can be
1) Extra lanes of motorway may not be required at peak due to the peak buses.
2) Car driversa get to work quicker due to the peak buses decluter ing the road
Quantif ing these is a bit of an art rather than a pure science. The costs still exist of course but if discussing funding peak buses these benefits have a place.
How to prevent peak only service in an extremely pendular urban rail?
My city is organized in such way that most passenger transport happens in peak hours of people going to their jobs in down-town and then coming back home at the end of the day. Down-town is mostly a commercial business non-residential area with mass transport coverage while the workers live at the periphery, in regions with minimal attraction, jobs and services, this means that most trains run crowded in downtown direction and empty towards the periphery, and the opposite at the end of the day: the system is clearly underused, trains are either absolutely crowded or absolutely empty.
What solutions there are to deal with this? I heard about cities elsewhere in the world employing an escalated system of working hours for public workers in order to distribute passenger dislocation through the day instead of concentrating it in an specific hour. But I never saw my city/metro/state governments venting such opinions and they seem oblivious to this.