Tag Archives: Amtrak

A Framework for Thinking About Long-Distance Travel

There’s been a lot of back and forth over the utility of Amtrak’s long-distance routes. Unlike regional services on the Northeast Corridor, these services don’t make enough money in fare revenue to cover their operating costs. To start, I think we need a good framework for looking at the alternatives, which as I see it, are as follows: do nothing, air, rail, bus, and car. Note that it is always important do include the do-nothing alternative, because transportation modes compete not only with each other but with the option of not making the trip at all.

To help compare these alternatives, I’m borrowing some graphical methodology from Cap’n Transit’s posts on commute options. Here are the five alternatives plotted with travel time versus cost for a representative trip, Chicago to Los Angeles.


And here they are with connectivity versus travel time.


I’m assigning an off-the-cuff assessment of connectivity rather than trying to come up with some numerical index. Feel free to debate the specifics but the relative rankings are sound. Doing nothing obviously gets you no connectivity. Your car takes you anywhere. Air offers very limited connectivity when it comes to intermediate destinations because, unlike a train or bus or car, the plane can’t make lots of stops en route. Bus and rail are somewhere between air and car. I’m putting bus above rail because bus currently serves more destinations, and because given the existing highway and rail networks in the United States, bus service could be scaled up more quickly. The difference in connectivity between modes is an important consideration for understanding which trips they are competing for.

Where does this leave us regarding long-distance Amtrak trains? For starters, I think it suggests that we cannot drive a huge modal shift from air to anything else other than doing nothing. Overwhelmingly, air and doing nothing compete only with each other. Air is far and away the fastest option for long distance travel, and it is not appreciably more expensive. The vast majority of air trips – business, personal, and freight – simply cannot tolerate having their travel time increased from 6 hours to 2 days. I live in LA, but I have a lot of family on the east coast. If it takes me 3 days to get there, I’m going to visit them less often. If we raise the cost of air travel, many trips will simply vanish.

Now, maybe you think a smaller total volume of cross country travel is a good thing. But that’s a different argument than shifting people from planes to trains, and you have to consider the potential for economic losses. That’s a topic for another post.

However, that analysis regarding air travel only holds up if air travel actually serves your origin and destination. As many have pointed out, lots of people riding long distance trains (and we can assume the same for bus) are not going from Chicago to Los Angeles. They are going from, say, Dodge City, KS to Winslow, AZ. Let’s rerun the time-cost analysis using those cities.


And that’s not even considering that the air option is probably subsidized by the Essential Air Service (EAS) program. We are clearly operating in a different space, and it’s easy to see why: air networks are fundamentally different than rail and bus networks. Serving large cities capitalizes on all the strengths of the air mode, while serving small cities highlights all the weaknesses. The opposite is true for bus and rail; serving small cities highlights their strength: the ability to aggregate trips with different origins and destinations onto the same trip. For these types of trips, air will never be able to compete with rail and bus (and car).

So, if the objective is to get fewer people to drive, the choice is between bus and train, and each offers different advantages. Rail offers a smoother, more comfortable ride, and it’s easier to get up and walk around the train. There is no question that riding on a train is a nicer experience. On the other hand, bus could immediately serve a set of origin-destination pairs that is an order of magnitude larger, and offer greater frequency.

My initial sense is that in the short-term, expanding bus service makes more sense in this case for one reason: infrastructure capital costs. Intercity bus service could be greatly increased without needing anything other more buses and perhaps increased bus terminal, layover, and maintenance space in some cities.

On the other hand, all of the long-distance rail routes are owned and operated by these guys, and I think it is indicative of transportation myopia on the part of a lot of transit activists that we’ve gone this far in the discussion of Amtrak and hardly even mentioned them. Any attempt at major expansion of long-distance passenger rail is going to result in the freight railroads demanding capital improvements. And why shouldn’t they? Here is an industry that, over the last 50 years, has been nearly strangled to death by overregulation and direct government subsidy of their competition, and not only have they survived, they’re killing it. They’re blowing trucking out of the water and dropping billions in private cash on capacity enhancements every year. If you’re a Class I RR, a money-losing passenger service is just another imposition by the federal government that hurts the viability of your business.

At this point, someone is going to note that both cars and bus service are subsidized by the provision of public highways and freeways. True enough. I see no technical reason that you couldn’t increase the cost of using highways to the point that passenger rail becomes profitable.

But if passenger rail becomes profitable, there’s no need for Amtrak anyway. The whole reason Amtrak exists is that passenger rail is not profitable. If passenger trains would make money and be a good use of track capacity, the Class I’s would start to run them. In fact, given how efficient the Class I’s are, they could probably reach profitability with lower ridership and/or lower fares than Amtrak. Tri-Rail doesn’t make any money but FEC seems to think they might be able to on a largely parallel facility.

So what’s the conclusion? I don’t know. Hey, no one said this was an easy question. I think it is clear that for long-distance travel that is not between two major cities, air isn’t going to cut it. In the short-term, increased intercity bus would seem to be the most readily achievable option. But there just aren’t that many trips being made between these minor city pairs anyway. According to the Kansas DOT, the AADT on US 56 west of Dodge City is 2,950. How many of those trips could possibly be served by long-distance transit in the first place? If the rural states want to pay for these services or have the feds chip in a little, I’m cool with that. If they don’t, I’m cool with that too. These trips are not the best type of trip to be served by transit and they are the type of trip where the weaknesses of cars matter the least. Transit advocates should have much bigger fish to fry than worrying about how people in Dodge City get around.

Note: travel times and cost based on Amtrak, Greyhound, and Kayak searches for a one-way trip leaving on May 15, 2013. For the small cities, I used Winslow for rail, Show Low for air, and Holbrook for bus (Amtrak serves only Winslow and Greyhound only Holbrook). I added an hour of travel time to get from Show Low to Winslow. The other airport option is Flagstaff and it is about the same distance/cost. For driving times, I used the Google Maps drive time and assumed two people driving 16 hours out of the day. For driving cost, I used $0.608/mile. You can tweak all of these assumptions; they are just rounding errors compared to the overall outcome.