Populist movements don't build themselves ...

... It doesn't matter what the "horse race" outcome of the campaign is, if we fight the campaign. Fighting it, we learn how to fight. Learning how to fight political battles, we become citizens again. Becoming citizens again, we reclaim the Republic that lies dormant beneath the bread and circuses of modern American society.

Sunday, November 29, 2009

Sunday Train: Frequency and Waiting on a Train

Burning the Midnight Oil for Living Energy Independence

I've been reading James McCommon's Waiting on a Train. And in cowed deference to the FCC, I will put the disclaimer up front that, yes!, I was more likely to read it and talk about it because Chelsea Green gave me a free review copy - since I would otherwise have had to wait until both it and I was in the library at the same time ...

... {of course, making me more likely to read it and talk about it is a gamble, since I'm not going to change my view of it because its a free copy - so if you have any publisher friends, warn them that if they reckon a book is a piece of garbage, they'd be better advised not to send a review copy}

The Chapter that is inspiring today's Sunday Train is "Amtrak Cascades: it's all about frequency". "Uznanski" is Ken Uznanski, former passenger rail chief of the Washington DoT:
"Once those intermodal trains can go through Stampede Pass, it will take some traffic off the main line and free up more room for additional passenger trains," said Uznanski.

By bringing the number of trains up to eight a day between Vancouver and Portland, ridership and ticket revenue will increase significantly. Currently ticket sales - what is known as farebox - cover 43% of the Amtrak Cascades' operating expenses; the state subsidizes the remainder. Run eight trains daily, however, the farebox recovery goes up to 70%.

It's all about frequency. When trains are frequency and convenient, ridership - particularly business travel - grows dramatically, said Uznanski.

It was a mantra I was to hear from experts all across the country - frequency builds ridership and only frequency significantly builds farebox recovery. Sure its great to have trains running more than 100mph in a corridor, but if there are only a couple of trains a day, they just aren't convenient enough to move people off the highway or away from the airport.

- John McCommons, Waiting on a Train, Chelsea Green Publishing: Vermont, p. 51


Back in Sunday Train: Rescuing the Innocent Amtrak Numbers from SubsidyScope, SubsidyScope's numbers proved to be hopelessly inept to serve as any basis for evaluating capital costs of individual Amtrak routes (even though SubsidyScope adopted the pretense that they could in fact be used for that purpose). However, as note, it might be possible to use their operating cost and revenue values as a rough guide to operating cost recovery.

Going back to the "Regional Corridor" services (shorter services running outside the Northeast Corridor), I tabulated a rough estimate of the services that had above median and below median operating cost recovery based on SubsidyScope's numbers.

So I've gone to the Amtrak timetables, and have done a rough estimate of effective frequency of weekday service. By "rough estimate" and "effective frequency", I mean:
  • If the corridor seems to be served primarily by a single service, I only count that service, while most of it is served by a long distance service as well, I include the long distance train along the corridor as a service
  • If there is a partial route that covers about half the corridor, I count it as half a service
  • Service three weekdays per week counts as 0.6 of a service.


Going back to that table, there is some direct confirmation of the argument. There is a median of three (3) services per weekday for all regional corridors services, with a median of four (4) services for those with above-median farebox recovery, and a median of two (2) services for those with below-median farebox recovery.

To get a visual overview, the Regional Corridor farebox recovery table was originally formatted with above-median recovery ratios in the left hand column, and below-median recovery ratios in the right hand column. So simply breaking it into two tables - corridors with three or more services per day and those with fewer than three per day - and just having a look. It confirms that, while frequency is not the sole and solitary factor driving farebox recovery, there is a tendency for more frequent trains to be in the left hand column and less frequent trains to be in the right hand column.

Three or more services per weekday:
Corridor Route Op Ratio Corridor Route Op Ratio
Corridor Services 58%
Chicago-St. Louis IL/MO 79% Albany-Toronto NY/ONT54%
Newport News DC/VA77% Capital CA 53%
Illinois-Saluki IL 70% Empire NY48%
Hiawatha IL/WI 66% Wolverine MI/IL39%
Downeaster ME/MA 64% Springfield Shuttle MA/CT32%
San Jaoquins CA61%
Keystone PA 61%
Cascades OR/BC 61%
Pacific Surfliner CA 59%
Pere Marquette MI/IL 59%


Fewer than Three Services Per Weekday
Corridor Route Op Ratio Corridor Route Op Ratio
Corridor Services 58%
Carolinian NY/NC71% Blue Water MI/IL 55%
Heartland TX/OK67% Zephyr IL54%
Ethan Allen NY/VT65% Vermonter VT/DC 54%
Adirondock NY/Quebec 59% Piedmont NC52%
Missouri River MO45%
Pennsylvanian PA/NY43%
Hoisier State IN/IL15%



Whole Systems Thinking

There are two really substantial competitive advantages to frequency - but both require more than a high frequency schedule alone to tap that potential.

The first benefit is meeting demands for time-of-day sensitive travelers. Some travelers have wide latitude in the timing of their departure and/or arrival - others are faced with a more rigid earliest departure time or latest arrival time. The greater the frequency of the service on a rail corridor, the more travelers are within the real potential transport market for the corridor.

But of course, frequent and (even worse) extended delays - which are a regular occurance in McCommon's rail travels - will drive away precisely those people most strongly attracted by high frequency services - the time-of-day sensitive traveler. If the main origin of a corridor in one direction has three services that are well time for morning, afternoon and early evening travelers, and a low percentage of trains arriving and departing on schedule with frequent multiple-hour delays, reducing the number of services delayed, and reducing the average delay for those that experience delays is critical.

On the other hand, the main source of delay is limited system capacity and aging infrastructure and equipment - and investments that substantially improve the capacity to operate on time tend to be the same ones required to permit additional services to operate. So "frequency" is to a certain extent a stand-ing for "reliability and timeliness".

The second big advantage is the market for day trips. If there is one service each way per day, then travelers on one or both sides of the corridor are excluded from using the train for day trips. But of course, the length of the trip itself is also critical for making a day trip possible - a six hour train ride with a departure at 6am and a return departure at 6pm makes a day trip possible in one direction (but not the other) ... but it won't attract many day travelers unless that is a compelling destination from that origin.

By contrast, if there are five to eight services each way per day and it is a trip of three hours or less, that is a corridor that can be in the frame for people taking a day trip.

There is an "exception that tests the rule" in the above tables. The Pere Marquette is one service from Grand Rapids to Chicago which is effectively one of five Amtrak services per day connecting Northern Indiana and Chicago (supplemented by additional metro rail services in northwestern Indiana) ... which leaves Grand Rapids about 7:30 in the morning for an arrival before 11am, and returns from Chicago about 5:20pm for a return before 11pm. So it comes in the frame for a "day trip to Chicago", even though between Grand Rapids and New Buffalo MI it operates as a single service each way per day.

Of course, its difficult to relocate more moderate sized cities to be a convenient distances from a sequence of other moderate sized cities or a major mega-regional center like Chicago. The way to bring more origin/destination pairs within a three hour time frame is with faster average trip speeds. The first threshold is the 110mph level, which with tilt-trains over a reasonably efficient alignment is time-competitive with cars, so many car-based day trips are also rail-based day trips.

The second threshold is the 125mph level - the speed of the first generation of bullet trains - which with tilt-trains over a reasonably efficient alignment results in trips noticeably faster than most cars. That means that there are trips that are solid day-trips by rail that are not day-trips by car, which offers another step up in competitive positioning.

At this point, someone will normally chime in that there are special benefits to using a car for a day trip - but that is just a one-eyed way of saying that different transport modes have different mixes of advantage and disadvantage. A major benefit of a train for a day trip is that you can relax on the way down and the way back, rather than having to work at chauffeuring yourself.

And of course, trip speed interacts with service reliability in service of frequency in a very basic way: the sooner you can be confident that the train will be at its destination, the earlier you can schedule the same train to run back the other direction. A five hour route can get at most three day services each way from a pair of trains - and the morning departure is likely to be earlier than ideal in order to fit in three. A three hour route can get four or five.

So in reality, improved trip speed, improved reliability, and increased service frequency are a complex, and thinking in a pure analytical manner about the effect of each as independent factors is a mistake - what is critical is the leverage that is gained when all three are improved as a system.


Service Frequency Relative to the Market

That is the main point that this section of the book brought to mind as I was reading it yesterday. However, there is also a point that comes to mind when looking at some of the low-frequency services with higher farebox cost recovery ratios.

How frequent is "frequent". When doing the median test above, there is a premise that three trains per day is three trains per day, in whatever market conditions. However, stating the assumption is enough to highlight that it is a shaky assumption.

The focus above was on the share of the traveling public that are "in the frame" for considering taking a train for their trip. However, clearly, the number of those "in the frame" that then chose to take the train depends critically on what the alternatives are. In relative terms, four corridor routes per day would be an extremely low frequency on the Northeast Corridor, where in the absence of regional corridor intercity rail travel there would still by very high frequency air shuttles and intercity buses in operation as well as shorter distance metro rail services that extend out to cover the shorter intercity trips.

Indeed, those four trains per day might be viable options for making a large number of trips, but if they are competing against a wide range of much more frequent alternatives, the number of people "in the frame" will only become obvious when planes are grounded or there is a closure of a major intercity road artery.

By contrast, four trains per day from Columbus to Cleveland and from Columbus to Cincinnati would be a substantially higher frequency relative to the market, because there are no metro rail services, very few intercity bus services, and a much more limited (and, of course, more expensive per mile) choice of flights.


Time to share your thoughts

Now's the time to share your thoughts on the future of rail travel in this country. Don't fret that you haven't read the book - or even if you don't want to talk about frequency, reliability and trip speed. Just say, "On a related point" and have at it - if its on trains, it'll count as a related point.

No comments: