RTA Bus vs. a Train Between Ann Arbor and Detroit? (Part 2)

One of my readers observed that she was a bit bummed out by the “yeah, but it simply can’t work for me” responses on one of my threads about the RTA’s recently proposed intercity bus service. Overall, there was a lot of ho-hum from the transit echo chamber. Response was positive, but perhaps a bit tepid. “Yes, we’d love this!” met the likes of “It’s still not a train!”

I mean. It’s true.

The Vernor bus in downtown Detroit.

Yea, we must face a sad truth. Amtrak, conditioned by decades of austerity from government bureaucrats who would rather erase it altogether, is unfortunately not an independently entrepreneurial entity.


In other words, Amtrak isn’t out there on the front lines hustling. But they do employ a lot of competent people. One of these is media man Marc Magliari, who is based in Chicago and spoke to us on the phone about the transit saga. Asked about the (almost mythologically) hypothetical high-speed rail link between Ann Arbor and Detroit, Magliari didn’t shoot it down. But he was careful to point out the distinction between high-speed infrequent service and lower-speed, frequent service.

“We grew ridership between Chicago and Milwaukee by 40,000 customers [per year], and that’s with no change in speed,” he said, referring to a service expansion on Amtrak’s Hiawatha line. “That’s just by running reliable service at a competitive fare that’s as good as driving, if not better.”

Fantasy transit maps for Southeast Michigan are a whole Reddit thang.

Frequency and reliability are the key here. There is no commutable rail connection between Ann Arbor and Detroit. Large portions of the trackage that does exist is in poor shape and it can’t accommodate travel faster than about 40mph. This means that it takes about an hour to get to Ann Arbor on the train, so a door-to-door trip is more like an hour and a half compared to an hour or less driving. 40 miles should take more like, uh, 20 minutes? 30 minutes with one stop, maybe?

Freight railroads, Amtrak, and certain state entities all separately own sections of track, so this is difficult to balance. Furthermore, it’s hard to get bureaucratic public entities together with for-profit corporations that are pretty abominable at doing anything other than caring about shareholder profits. I’ve always been able to reach Amtrak’s media people. I cannot say as much for the Class I railroads. Telling!

The same federal government that granted railroads free land during the pillaging of the great American West in the 19th century is now getting shafted in favor of those railroads’ shareholders. Figures, I guess.

Michigan Central Station in Corktown, now owned by Ford. The tracks pictured run under the Detroit river through an old tunnel that, our Canadian train history aficionado tell us, is in pretty rough shape.


A locomotive engine costs a few million dollars. And passenger cars might also cost, well, probably also few million dollars— each. Fast, frequent service between Detroit and Ann Arbor would likely also best run on electrified trackage.

An extremely nerdy post from rail and transit wonk Mike Hicks points out that long trains aren’t necessarily more cost-effective at carrying passengers. But Hicks notes the cost per seat of a Boeing 767 is far more expensive than a seat on a train in terms of total capex. Like, ten times more expensive. And this is to say naught of the outrageously high costs of maintaining an airport, which requires 24/7/365 space conditioning, grounds and tarmac maintenance, a mess of security, and providing a spendy stream of fuel to jet engines. Hicks wrote in 2015:

“A Boeing 767 with 350 seats runs about $180 million, more than $500,000 per seat. In contrast, a fairly standard train with four 90-seat Superliner-style bilevel cars (360 seats total) and a new locomotive would probably run $12 to $18 million, up to about $50,000 per seat.”

There’s no holy grail of costing here, because governments decide to build highways all the time without thinking about what effects it will have. One thing is certain, though, and that’s that the ability to cop a 10% market share of the aforementioned 215,000+ car trips per day between Ann Arbor and Detroit would easily fill up a bunch of trains. Given that transit modeshare even in Los Angeles exceeds 10%, that’s a safe, conservative estimate. Most of these trips are single passenger trips, and the ones that aren’t are likely balanced out substantially by the percentage of the 215,000 that are trucks, not cars.


Hicks’ example train fits 360 people. That would work out to 56 trains to carry those 20,000 people. Sounds like a lot until you do the math and figure that 22-hour daily service with trains leaving every 15 minutes would be 88 trains per day, since we’re assuming that demand isn’t normalized hourly. 88 trainsets, based on Hicks’ metrics, would cost $1.06-1.58 billion to buy. At the conservative end of his cost estimates, running this service would cost $88,352 per hour, $1.9 million per day, or $709 million per year.

Sound like a lot of money? It is may well be. But 215,000 cars cost drivers an average of $1.9 billion per year to own, nearly three times the cost of operating the service. Add to that hundreds of millions of dollars in costs of highway maintenance. Add tens of millions more for the costs of congestion, air pollution, and car accidents. Fare recovery calculations necessarily miss this for the purpose of comparing transit to roads because of how cleverly highway spending externalizes so many costs.

However you slice it, gas tax and user fees come nowhere near making up the total cost of roads. Indeed, gas taxes in many jurisdictions are used to fund things other than roads, with notable exceptions in states like Ohio and Indiana that enjoy toll roads as a major source of revenue.

Amtrak’s Marc Magliari notes that it’s possible from a funding standpoint because only 15% of Amtrak’s funding is federal. Amtrak’s limited funding pool is largely responsible for its inability to entrepreneurially develop new projects in individual states. Rather, Magliari says, that push has to come from the states, probably the department of transportation, which contracts Amtrak to operate rail service.

“If MDOT came to us and said, ‘hey, we want you to run trains between point A and B,’ we’d certainly look at that,” he said.

The takeaway from this napkin math is that train service is doable. But it’s possible that it will require that bus service be a smash hit. First things first.

All aboard?

Nat Zorach

Nat M. Zorach, AICP is a city planner, community development professional, and MBA candidate at American University's Kogod School of Business, based in Detroit.

Leave a Reply