Last week I wrote about the tension that the US national rail carrier Amtrak faces between ridership goals — which require focusing on its best markets — and coverage goals — which require covering the entire country. I was applying a framework that I developed for urban public transit, but that seemed relevant enough to be useful in discussing Amtrak. The post attracted a great comment thread with lots of lively disagreement. So let me briefly respond to the commenters’ main objections. To be clear, though, I am not an intercity rail expert, so I am intentionally keeping a high altitude on an issue with lots of rich detail. I may be wrong, and if so I’m sure commenters will say so, with receipts.
To recap: Amtrak sees its biggest growth opportunities as corridors linking major cities under 500 miles. (This is also the ideal distance for high-speed rail, but Amtrak’s plan is about providing medium-speed service even as high-speed rail is developed in parallel.) The company’s new vision document, Amtrak Connects US, is almost entirely about investing in such corridors, serving trips lasting a few hours at most. But the US also has huge states with small populations, and there, Amtrak runs its multi-day long-haul trains. Worldwide, only Russia, China, and Canada have longer passenger train lines than the US. The Chicago-Los Angeles train formed by the Texas Eagle and Sunset Limited, is almost three times longer (in both time and distance) than the longest line in Europe.
Amtrak’s long-haul trains have two roles: they are marketed as a scenic and relaxing way to travel between cities if you’re not in a hurry, but people also use them for access to towns along the way. (I was thinking of the first market when I carelessly called them “land cruises, and the objections of many rural readers is understandable.) These services do as well as they do only because they combine both markets , so each kind of user depends on the other for the total package to make sense. You can see the “land cruise” role competing with the local access role in the service design. For example, these trains sometimes follow routings selected for scenery but not population, most obviously in Montana where the Empire Builder‘s far-northerly routing goes through spectacular Glacier National Park but misses all of the state’s largest population centers.
Many commenters objected that these can’t be coverage service because their ridership is actually pretty high, so I should be clear that by ridership I always ultimately mean ridership divided by operating cost. These trains are expensive to run. I will not delve into the arguments over Amtrak’s accounting for these trains, but it’s easy to see that they can’t be using labor very efficiently. The many rural stations need staff, all of them serving just one train at a time. The long runs require crew to rotate in and out of duty on board, so those people need places to sleep. The frequent delays mean that employees (both on the train and at the stations) often work longer than their scheduled shifts, incurring overtime. It’s all very labor intensive, compared to short train lines where you turn over passengers quickly and most your staff can go home every night.
Meanwhile, the fares, if you don’t get a private room, are almost zero. Portland, Oregon to Wisconsin Dells in central Wisconsin, coach, nonrefundable, a month out, is currently showing a fare of only $128, which is about 6.5 cents per mile. A similar coach seat on the Northeast Regional train from Washington DC to New York with the same features is $31 or 13 cents per mile, twice the cost for a service that almost certainly costs Amtrak less to provide on a per passenger basis. So the cost of these services lies not just in their inefficiency but in their low per-mile fare, which of course is what keeps them affordable enough to have a lot of use.
So can these services be justified by their ridership? It depends entirely on whether you count ridership in passengers or passenger-miles. If you want long-distance services to look good, you count passenger-miles, which values one person going ten miles as much as ten people going one mile. In the urban transit context, there is a similar suburb-vs city debate. If you want to make long-haul suburb-to-city service look better than shorter runs within the dense city, you cite passenger-miles, but some inner city folks will object that those long distances are a side effect of poor land use planning and location choices, not an actual benefit. The benefit lies in how many actual human beings have gotten to where they’re going. There’s no objective way to decide the question between measuring passengers and measuring passenger-miles. It comes down to what kinds of trips you value. A rural-urban divide is unavoidable on that point, because long distances are the defining fact of rural life.
My other question about the “rural lifeline” case for the long-distance trains is: Do these services really match the pattern of rural demand? Probably not. If you were optimizing a network for access to rural communities, it wouldn’t look much like these long-haul lines. Instead, it would look like a robust set of links between smaller centers and nearby larger ones. Most rural intercity travel demand isn’t for the thousand-mile trips that Amtrak makes. It’s more local, under 200 miles, with smaller centers usually needing access to larger ones that have more jobs, essential services, and onward transportation options.
This, of course, is the role of a robust bus system, working with trains as appropriate. The major Australian states, for example, have publicly funded intercity bus+rail systems that connect most rural towns into their nearest larger center, with onward connections, bus or rail, to the big city. This pattern is most likely to connect small towns to the destinations most useful to them. As commercial intercity bus services wither in the US, there will be an increasing role for state funding and management of similar services. I’ve beaten this drum before, but right now, in the US, rural transit between towns is mostly provided by county-level agencies with minimal state coordination. Every county line is a potential barrier for travelers on those services. At best adjacent counties have to make deals to allow service to flow over county lines to serve the trips people need. What’s missing is often a network managed and funded by the state, running many of the lines that commercial carriers once ran. Pieces of that might be rail, but a lot of it is probably logical bus networks that may cross the state but don’t cross half the country.
Does that mean the long-haul trains shouldn’t exist? Not necessarily. Personally, I’d love to see more of them. Overnight trains are having a renaissance in Europe, and this suggests new markets where some useful rural service could be combined with overnights between bigger cities.
But still, Amtrak is right not to focus on these trains as part of a ridership-maximizing strategy. Right now they are part of how Amtrak meets its competing goal of covering the whole country. But because this has become an exercise in counting states, and thus Senate votes, the result is such thin coverage that most rural communities don’t have access to it. Even towns with stations may not have service in the right direction, or at a useful time of day, or with the decent on-time performance that comes only from shorter lines. Most of the lifeline access that the rural US needs is logically provided by robust bus networks, managed by states if they can’t make a profit. By all means let’s keep and celebrate the long-haul trains, and expand them as appropriate. But let’s not let them distract us from what it would really take to provide liberating transportation to rural America.