America is about Freedom. Freedom comes from Frequency.

Greetings from East Glacier, Montana! I hope everyone back in Greater Philadelphia had a good holiday weekend, and that the weather I hear you’ve been having hasn’t been inconveniencing you too badly. The scenery here is breathtaking, and the weather has been mild and cooperative.

I want to tell a few stories about Frequency Anxiety and its opposite, Transit Freedom. It’s been the theme of this week so far.

The first story takes place on Amtrak. I woke up on the Capitol Limited with a bit of a rash on my neck, a bit irritated, but nothing major. Thinking it a reaction to the strong disinfectants used on the bedding in sleeper cars, I resolved to treat it with antihistamines and get on with the trip. The Cap arrived at Chicago Union Station only about 90 minutes late, and after lunch with friends in Chicago, we made our connection to the Empire Builder. I went to sleep somewhere along the Mississippi River south of St. Paul.

When I woke up near Minot, North Dakota, the rash had spread, and I was in agony. As my wife was snapping photos of my neck to send to medical professionals back in Philly, I was looking up medical care options along the route. There is basically nothing in the way of health care infrastructure in East Glacier, with the nearest doctor’s office in Browning, and the westernmost full-service hospital along the Empire Builder route before reaching East Glacier is in Havre, 4h16m and 174 miles away. Fortunately, the word back from my doctor was that I didn’t need hospitalization, but I did need prescription meds. The nearest contactable pharmacies to East Glacier were in Cut Bank, an hour east and the previous stop, and Whitefish, three stops west, but a similar driving distance.

Of course, with evening pressing in, I couldn’t just get off in Cut Bank, even if I was guaranteed to get across the small town to the pharmacy before it closed. The Empire Builder only runs once a day, and with no way for me to self-transport between Cut Bank and East Glacier until the next day’s Empire Builder, we rode through to East Glacier, rented a car, and drove to Cut Bank the next morning. Had there been another Amtrak schedule available in that 24 hours, getting off the train might have been a more viable option. (For what it’s worth, my case of shingles has been comparatively very mild, quite possibly due to timely treatment. Still, please, vaccinate your kids against chicken pox, because this has been pretty damn miserable.)

We only rented that car for one day. Our primary plan for the week was to rely on shuttle buses from East Glacier to trailheads along Going-To-The-Sun Road via connection at the St. Mary Visitor Center. This turned out to be a terrible plan; you can count the buses to or from East Glacier in a day on one hand, and the Eastside shuttle between St. Mary and the Continental Divide at Logan Pass only runs every 45-60 minutes, and with frequent random delays due to the roadwork. We spent as much time waiting for buses as hiking. And the uncertainty in trip times and the paucity of connections meant that we had to make our problem worse by building in large amounts of buffer time to ensure our connections. We were back on the phone with the car rental company before we even left St. Mary.

In both cases, the transit system that we wanted to use, and might have served us, couldn’t because it ran too infrequently, and we had to fall back on an expensive car rental to bail ourselves out of the jam.

The third story is about wider effects, and not just our anecdotal experience this week. It’s this post that Matt Yglesias apparently phoned in to Vox, about the distribution of Amtrak ridership. Now the fact in the headline is unquestionably true: half of the passengers that stepped onto an Amtrak train in 2013 did so at New York Penn Station, Washington Union Station, or our own 30th Street Station. So far, so obvious. But Yglesias turns that into a conclusion that Amtrak’s “geographically expansive service” is “smart politics … but it makes very little business or transportation sense.” This is only true if you see Amtrak’s one-a-day service on its long-distance service as inevitable. The difference can be seen in the states where there are state-sponsored corridor trains complementing the national long-distance network: California, North Carolina, Upstate New York, Illinois, Virginia, Michigan. Denver is the same size as Seattle, but with only one train east and one train west, to Seattle’s five trains south, two trains north, and one train east, the aspiring transit capital of the west can only scrape up a sixth of Seattle’s ridership. Denver finds itself lagging much smaller cities like Syracuse, which has four daily trains, and sends about half of its ridership to the New York City area, six hours away.

Now, is all of the anemic ridership at Amtrak stations along long-distance routes due to Frequency Anxiety? Of course not. Nor is there a good case to be made for blindly increasing frequency on long distance trains, as long as we don’t have a plan to fix its most glaring problems, like the financial hemorrhaging of the diner car service. But on routes with clear latent demand, like the Empire Builder, the Lake Shore Limited, and the Capitol Limited, Amtrak should clearly be thinking about expanding frequencies along the entire route. At least on some of these routes, I firmly believe that increased revenue from ticket sales will grow even faster than the near-doubling in cost. Additionally, Amtrak should increase the number of corridor trains or point-to-point sleepers that supplement and support long-distance trains on short segments, like Washington-Lynchburg for the Crescent, Chicago-St. Paul on the Empire Builder, and Oakland-Sacramento-Reno on the California Zephyr.

And closer to home, these stories are strong arguments for increasing service on the Pennsylvanian, as has been discussed here before. The argument for increasing service on low-frequency commuter rail lines where ridership is being suppressed by an obviously sparse schedule, like the Atlantic City Line and the Wilmington/Newark Line beyond Marcus Hook, is closely related. But mostly, it’s just a recapitulation of what we should all know by now: Transit, on the block scale and the continent scale and everything in-between, is about our freedom. Hobson’s Choice is not freedom. We need to encourage the growth of freedom by increasing the usefulness of transit in smart ways. The best and easiest way is to grow frequency.

We’ll be back in the Cradle of Liberty next week.

Join the Conversation


  1. Excellent post. Hope you’re feeling better and the rest of your trip is great. Definitely agree about transit frequency, and the freedom it gives you, being key to its success. That’s why I fell in love with it in the first place. Most people don’t even look at the PATCO schedule before they get to the station in the morning, they jut show up and get on a train. And it’s 24/7. Compare the freedom that gives people versus the infrequency and headaches that hourly SEPTA regional service or even worse A/C and Wilmington service, like you mentioned.

    Also, as someone who recently took a long(ish) distance train for the first time, I’d really like to see them bolster the lines that make sense. I took the Adirondack train to Montreal a month back, and it was a great trip aside from the fact that track conditions make it so much closer than driving. I still preferred it to driving, and there were tons of people on the train, but you can imagine it being even more popular, possibly justifying more than once-daily service, if they improved the track to get you up there in 6 or 7 hours instead of 11.

    1. Thanks! Get to wake up tomorrow for an extra-early Empire Builder, courtesy of BNSF and the Bakken Shale.

      It’s gotten to the point where I feel weird posting anything without a citation to Jarrett Walker, so here’s one: have you seen this video of a presentation he gave in Toronto recently? I linked it somewhere to some people, don’t remember if you were among them, and my internet really is dialup-quality, so I can’t check. If you haven’t read his book, it’s the best distillation of his ideas that I’ve seen. (OTOH, if you watch that video and read his book, my blog might be a lot less interesting…)

  2. I disagree with your assertion that there is sufficient latent demand for higher frequency on the long distance routes. If the demand existed, ticket prices would be higher, and Amtrak would be closer to profitability on those routes.

    For your specific example of Denver versus Seattle, I think it’s hard to justify more service that connects Denver, because despite the metro area having a similar population, there is less demand in Denver for this type of service. The proportion of Seattle households without a car is much higher than in Denver, and the closest cities to Denver (Salt Lake City and Albuquerque) both have low proportions of carless households, and much lower populations than the closest cities to Seattle (Portland and Vancouver). I haven’t been to Denver, so please correct me if I’m wrong about this, but I would guess that the walkability of the area around the Denver Amtrak station is probably much less than Seattle.

    Now this is not to say that we should not be subsidizing long distance rail service in the west. But considering that the latent demand is so much higher in the northeast, the subsidies should go there first. Particularly on the New York to Washington route, which is in so much demand, that even with the highest frequencies in the Amtrak system, they still turn a profit.

    I mean, I used to live on the 1 in New York. I now live off the Broad Street Line, and I would like it if the frequency was more like the 1. But there’s much, much more demand for the 1 because there are many more people living near the 1, and those people are much less likely to have a car. Even during peak times, I can always find a seat on the BSL, where I usually could not on the 1, even with double the cars per train.

    I guess what I am arguing is that the demand is largely independent from the frequency. The BSL is actually pretty nice as transit goes, but hardly anyone lives near it. If the areas adjacent to the BSL were developed like the areas adjacent to the 1 (virtually no off-street parking, 6+ stories of housing instead of 2 or 3), then it would have more comparable demand. And you could say much the same for Denver versus Seattle, or for the Empire Builder versus the Northeast Regional. The Empire Builder sometimes sells out, but it still loses money, and the Northeast Regional doesn’t lose money.

    1. My assertion that there is latent demand has more to do with schedule and convenience, not pricing.

      For a clearer example, take North Dakota. The highest ridership station in the state is Williston. There’s a lot driving Williston traffic, between the oil patch workers and the lack of other options, but Fargo is still a much larger city and a shorter ride to the major destinations of MSP and CHI, and should have higher traffic. It doesn’t, because the Empire Builder stops at Fargo in the middle of the night. Unless you’re an insomniac or under duress, Fargo might as well not be an Amtrak station.

      That’s an obvious case. Less obvious is just convenient timings. People going from, say, Chicago to MSP, might not be able to leave in the afternoon because getting in at midnight is too late, but might be willing to leave in the morning to arrive in the early evening, or leave in the late evening to arrive at dawn. But because Amtrak doesn’t offer them a conveniently timed trip, they fly, or drive. (This may be me if I go to the Strong Towns National Gathering in September.) Don’t forget to *repeat* this analysis for the MSP-CHI leg of the round-trip.

      And ultimately, increasing frequencies doesn’t scale costs linearly. The cost of keeping up every station building, commissary, switcher and crew, etc. between Milwaukee and Bellingham or Vancouver, is charged to the account of the Empire Builder’s one train a day in revenue. That’s a lot of capital cost, and not a penny more would be needed if a second train ran nine hours behind the current train. None of those are as expensive as Penn Station, but Penn Station has 80+ Amtrak trains per day, plus LIRR and NJT as tenants, so the cost gets spread out a little more.

      I will address Denver and other topics more directly, but I am typing this on the WP app on my phone from somewhere near Wolf Point, MT, and signal is spotty, so it should wait for a second comment.

    2. The LoDo area around Denver Union Station has been transformed over the last twenty years into a great walkable neighborhood, and over the next ten years it will become the hub of Denver’s epic crash-buildout of rail transit known as FasTracks. (No link because phone.) It should have roughly the same sort of catchment as an origin and a destination as Seattle’s King Street Station.

      Again, as with the CHI-MSP example, there is more than one way to handle the distance between Denver and Chicago. One way is what is done now: Chicago-Omaha in daylight, Omaha-Denver overnight. Just as valid (especially because the CZ is routed around the main cities in Iowa) is Chicago-Omaha overnight, and Omaha-Denver in daylight. With a second train, you can do both. (The main cities of the CZ do not line up quite as nicely as the EB for non-overnight timings on an eight hour offset, but it’s close.). That and Colorado finally biting the bullet on the often-discussed Front Range Corridor, which I will go out on a limb and say will happen in the next five years as equipment becomes available and BNSF prepares to abandon Raton Pass, and then Denver has no excuses.

      The Northeast Corridor lost money until very recently, and depending on how you calculate the capital costs, still loses money. I am not going to throw that stone at the west from my northeastern glass house, and I encourage everyone who comments here to do the same, although it’s not and won’t be an enforceable rule. And my main point was about about routes where there is only one, or a few, options per day; the comparison to routes with 21 or 49 trains a day is a slippery one, and prone to fallacy. I would have respected Yglesias’s piece more had he actually determined how many Amtrak passengers never left the NYP-WAS segment of the NEC; I suspect it’s far, far lower than half the total.

  3. Supposing there is more demand on long distance routes, and supposing demand would double or more while costs almost doubled, I imagine that on most routes Amtrak would still lose more money. Less money per rider, but more money total. Given the current budget climate and the service/infrastructure badly needed on core routes, I think that makes increased service a non-starter.

  4. Great post. I’ve thought a lot about how often counts as “frequent” based on trip length, and here’s a post I wrote last year:

    Obviously this gets tricky with longer-distance trains that could be serving people traveling a variety of distance lengths, although I think the principles I outline still hold there for individual passengers — i.e. once-daily service is much more likely to satisfy someone going from Seattle to Chicago than Seattle to Spokane.

Leave a comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: