Fingers Crossed For Autonomous Inter-City Shuttles
I just saw this news story announcing that Amtrak is halting service between Chicago and Indianapolis after the state of Indiana decided not to continue funding the service.
It was timely and personal to me because I was on the Megabus from Chicago to Indianapolis when I wrote this post. While the news is a bummer for those interested in transit options, the service was already so infrequent and slow as to make it impractical — it runs 4 times a week and takes about 5 hours door to door. It is so infrequent that when I searched for transit options on Wanderu (which aggregates Amtrak, Megabus, Greyhound, and other transit services), I didn’t even see it as an option!
Every time I take the bus between cities, my reaction is ‘I’m glad I didn’t drive, but I wish I could have paid more for better service.’ The bus is fine, although frequently either the wifi or power will not be working, and the ‘bus stop’ is a curb across the street from a strip club.
I don’t own a car myself, but I do a decent amount of travel to different cities in the Midwest to identify new startups and partners, so I’m always looking for different travel options. Once I reach my destination, there are plenty of options to get around (lyft, uber, walking, public transit, bike share, scooters, etc…) — it’s really the city-to-city transit that is underserved.
This is one more reason I’m excited about the potential for autonomous vehicles- particularly minibusses that can drive small groups of folks a couple hundred miles to different urban areas. For context, I wrote this from a full-size coach bus, but there were all of 11 paying passengers on the bus including me. It’s not hard to imagine 1 or 2 autonomous shuttles servicing a very similar route. I’ve included some back of the envelope calculations below on the potential economics of such a hypothetical shuttle for those interested, which suggests that, at a minimum, this type of approach could be a very attractive transit solution(*).
While long-term, this should be no-sweat, I’m curious about what solutions will develop in the next several years. For example, if we have the autonomous functionality to handle highway-only driving, what does that mean for new transit options and infrastructure? Do we need staging areas immediately adjacent to the highway that are geo-fenced for autonomous shuttles? For example, my bus picked me up 3 blocks from the highway — would that be close enough? And will some land that is currently very low value because of its proximity to a noisy interstate suddenly become much more in-demand? I don’t know the answers to these questions, but I’m very interested in seeing where they lead, and in any new inter-city transit option that becomes available.
_____________________________________________
(*) Back of the envelope service, revenue, and expense assumptions:
Expenses: Pulling numbers out of a hat, I can imagine an autonomous mini-shuttle costing in the $150,000-$200,000 range (or what you’d pay for a couple of Tesla Model S vehicles; for context, a diesel bus is around $500,000 while a Proterra electric bus is around $750,000). Let’s assume we purchase 5 shuttles (2 for each direction plus a spare) — we’d be looking at around $1.25 million of capital investment. Let’s generously assume $200,000 in overhead (software, insurance, local staff, electricity) per year; even if you bought new vehicles every year (unnecessary, but let’s be conservative since I am probably ignoring some other cost), you’d be looking at a bill under $1.5 million per year.
Service: What would that get us? Well, before charging anyone for tickets, let’s look at what service has looked like historically. The Indy-Chicago line averaged 30,000 rides over the last 2 years, or 120 per weekday (call it 60 each direction).
If each shuttle runs 3 times a day (morning, mid-day, and evening, with some breaks to charge batteries) and takes about 10 passengers per trip, then you could on average move all of the passengers that ride on that Amtrak service with just a handful of vehicles.
So, before you charge anyone for tickets, you could theoretically increase the availability of service by an order of magnitude (from 4 days a week to 3 times a day), at less than half the cost of subsidizing the Amtrak route ($3 million per year, or $100 per ride).
Revenues: a shuttle has 10 passengers paying $25 a person (Megabus rates), each shuttle runs 3 times a day, 5 days a week, every week. A shuttle utilized at this rate would gross just shy of $200,000 in revenue a year. At Amtrak ticket prices (averaging $33, based on the reported information in the article), that would be more than $250,000 in gross revenue per shuttle. At the level of the system we looked at for expenses above (5 vehicles, 1 not in use), this hypothetical service would bring in $1 million a year in revenues and cost $1.5 million a year to operate. If on the other hand, vehicles were good for 2 years, then you’d be looking at annual expenses around $700,000 against $1 million in revenues).
B.O.T.E. Conclusions:
Is this a cash cow business? It’s not obvious that it would be, although this is based on a very rudimentary analysis, so take the conclusions with a grain of salt.
That said, the economics appear dramatically more attractive than that of traditional (ie slow)inter-city train service in the US, which requires heavy subsidies. While longer routes are still definitely handled better by investments in fixed infrastructure that would allow for higher speeds, there are a number of shorter routes which could still be competitive transit options using existing highway infrastructure. Importantly, transit service provided by autonomous shuttles (via existing roads) requires very little infrastructure and long-term planning to implement, enabling different areas to pilot, implement, and adjust these sorts of services with limited financial obstacles. This should significantly lower some of the barriers to implementing these new types of services.
