Superforecasting/Don't Believe the Hype and 1 million robo taxis, Uber/Lyft trips 1/5 the current cost, and the Boring Company's tunnels or "Cities and mass transit versus personal transit"

At a Little Free Library I came across an "Instaread" (like the back in the day "Cliffs Notes" on the book Superforecasting by Tetlock and Gardner, which was quite interesting.  While I wouldn't call myself a superforecaster, I do constantly seek new information and experiences to better hone my analysis and understandings as they relate to transportation and urban revitalization.

Coming across some recent "forecasts" by businesses and stock analysts and the general hype that has accompanied them for some time, I just can't believe how much money must be sloshing around in the capital markets, to fund concepts that are not necessarily nonsensical, but are disconnected from "the market" or "how things work" in practice.



As far as cities and transportation go, there are three elements to keep in mind.

First, the road network (miles of roadway and number of lanes) is pretty much fixed (unless you add tunnels).  In turn this puts a limit on the number of vehicles which can be accommodated.  (Zurich is one of the few cities that takes this to the logical endpoint, by "metering" how many vehicles can enter the city.  See , which means that there is only so much capacity for vehicles.

The advantage of a (40 foot_ bus over a car is that while it takes up the space of three cars it carries up to 60+ people while the three cars carry 3-5 people.  Heavy rail does even better in dedicated above-ground or underground service, carrying tens of thousands of people per hour.

Various images produced that show people throughput by mode illustrate the point.


This graphic shows corridors, so the "mixed traffic" mode it is referring to is freeway driving, which doesn't correspond to driving within cities, where lane capacity ranges from 600 to 1,500 cars per hour (the latter on fast moving very wide arterials and one way streets), less than a freeway, because there are many more intersections crossing the street.


Second, limited road capacity means that the more cars you have, the fewer number of people you can move through a constrained system.

Ride hailing doesn't substitute for a transit vehicle, which can carry 60 or more people at one time, it substitutes from a single occupant automobile vehicle.  It doesn't improve throughput even if it reduces the number of owned cars and the number of cars that are parking in on-street or off-street parking.

The third element concerns market demand, and whether or not there is enough demand to sustain a working business model. Transit/shared vehicle users are clustered in the largest center cities. Most of the rest of the US is car dependent, and people drive a lot and long distances. To get to the point of sustainable market demand may not happen in my lifetime, and hopefully I still have 30+ years to live.

In the book Diffusion of Innovations, Everett Rogers explained the "take up" of new technologies and ways of doing things. It takes some time before the amount of use reaches level that make business operations economic.


Building on this work, Geoffrey Moore has written many books on the trough in business after innovators and early adopters have taken up new technologies, and before late adopters. In the book Crossing the Chasm: Marketing and Selling High-Tech Products to Mainstream Customers he terms this trough, "the chasm." From Wikipedia:

Moore begins with the diffusion of innovations theory from Everett Rogers, and argues there is a chasm between the early adopters of the product (the technology enthusiasts and visionaries) and the early majority (the pragmatists). Moore believes visionaries and pragmatists have very different expectations, and he attempts to explore those differences and suggest techniques to successfully cross the "chasm," including choosing a target market, understanding the whole product concept, positioning the product, building a marketing strategy, choosing the most appropriate distribution channel and pricing.

Image from "Models for Predicting the Future: Geoffrey Moore’s 'Crossing the Chasm'”, Smith House Design.

"Visionaries" include financiers and stock market analysts, and like with Chinese free floating bike share, scooters, and the like, they are "touts," not necessarily independent analysts.

While concepts such as shifting from ownership to fractional use have been around for awhile (e.g., "timeshares") and in sectors now ranging from clothes to travel stays to car sharing substituting for car ownership and taxis, the idea is that the cost of fractional use renting should be less than what one pays to own and/or operate and maintain the items. For some of these markets, they don't need a "mass market," because they can charge premium prices.

Given the present cost structure, for ride hailing to work, people will have to be willing to pay more than the alternatives (transit, car ownership).  That's not how you make a mass market.

-- "Disrupting The Car: How Shared Cars, Bikes, & Scooters Are Reshaping Transportation And Cannibalizing Car Ownership," CB Insights

Note that I think the metrics are off in the CB Insights analysis, but their segmentation of the mobility market by trip length is the way to go. And the headline should be "Disrupting The Car... [in a handful of cities]."

Note that as the cost to purchase an automobile continues to rise, and if capital supporting car leasing shrinks somewhat, the cost curve can shift, for some segments of the market, to ride hailing and other alternatives.


Maven ad in a DC bus shelter.

In the meantime, Uber and Lyft don't forecast moving to profitable operations anytime soon and car sharing companies are merging (BMW and Daimer now jointly operate ShareNow) and GM's Maven has just downsized significantly ("GM's Maven exits show tough road for mobility," Automotive News). From the article:

General Motors and other automakers envision bringing in significant profits from alternative ownership models such as car-sharing and subscription programs. Someday.

But for now, such mobility services are generally big money losers. And executives are increasingly discovering that they don't have the stomachs to let those businesses hemorrhage so much cash while they wait for technology and demand to reach the point where the services help the bottom line instead of hurt it.

GM, after expanding its Maven mobility brand to 17 metropolitan markets in the U.S. and Canada since January 2016, last week announced a "shift in strategy" that included exiting eight U.S. cities to concentrate on areas with "the strongest current demand and growth potential," the company said.

An op-ed in the Financial Times jokingly described venture capitalists funding Uber and Lyft as "transportation philanthropists" since each trip is heavily subsidized and users don't pay the real cost--irrespective of the constant pressure on reducing driver earnings to lower costs. (Although note, Maven is a late entrant to the market.)

Extra credit.  These initiatives are still focusing on stoking automobile sales, not in shifting people to more optimal mobility ("Further updates to Sustainable Mobility Platform Approach" and "DC as a market leader in Mobility as a Service (MaaS") choices.

=======

Tesla's one million robo-taxis by 2020

Obviously, Elon Musk has to keep up interest in Tesla to stoke the stock. But that doesn't make a market.

Tesla Model 3 in a city setting.

This has two elements.  The first is demand for "taxis/ride hailing vehicles."  Data I saw from the University of Minnesota said that the base cost for an automated vehicle are likely to be $1 per mile, which doesn't include a profit margin for the operator.

Given the type and number of trips people make and the cost of gas, most people aren't going to find switching from owning a vehicle to paying per use to be cost effective, especially given the average car is driven 12,000 miles/year, and the cost to own and operate a vehicle is about $9,000 year. 

Longer suburban trips are going to be more expensive. And cities don't have the road capacity to sustain a switch from transit to the car. As Jane Jacobs said once when asked 'why aren't there enough roads?', she responded "you're asking the wrong question. The right question is why are there so many cars?"

The second is the ability of autonomous vehicles to be used legally.  While the demand "for taxi rides" is in the city, the easiest place for AVs to work is on limited access freeways. But AVs are far away from being approved for use in mixed traffic.

UBS says the cost of ride hailing trips will plummet by 80% 

According to Business Insider, investment bank UBS, in a research report, claims that in 10 years, ride hailing trip costs will decline by 80% because of robocars, making the trip cost competitive with transit.

I think the important thing is to distinguish between variable trip costs and the additional cost of buying/leasing and maintaining the vehicle.

This article, "Cost-based analysis of autonomous mobility services," from Transport Policy (64:1 2018) argues that driver wages makes up 88% of the cost of a typical trip.  Although these costs don't include the overhead of "access to app" charges and profit margins.

And the current model, relying on independent contractors to provide vehicles, offloads the cost of maintaining large vehicle fleets to the drivers.

Irrespective of the cost per trip dropping or not, as discussed above, the transportation system in cities doesn't have the capacity to expand to accommodate significantly more trips by motor vehicle.  Taking on that burden will be significant.

But the UBS simulations didn't consider another factor: per trip "sales" taxes.

For example, to encourage sustainably modes, gasoline excise taxes are very high in European countries, and the cost of a gallon of gas ranges from $6 to $8 dollars.

Slap a $5 or more ride tax on what would be a $2 ride hailing trip as a transportation system management fee and we have little to fear.

Boring Company Tunnels

While I do think we need to consider underground tunnels as an option in more places, as a way to shift commuter traffic from city-serving streets:

-- "Tunnelized road projects for DC and the Carmel Tunnel, Haifa, Israel example--tolls," 2011
-- "London Mayor proposes roadway tunnels to divert surface motor vehicle traffic and congestion." 2016
-- "Maryland HOT lane study versus "corridor management" and regional scaled transportation planning," 2018
-- "Who knew? There's been a freeway deck in Oak Park, Michigan over I-696 for almost 30 years," 2018

I doubt that Elon Musk's Boring Company has come up with a way to build tunnels more cheaply, as transportation officials from Virginia recently confirmed ("Virginia transit officials drove through Elon Musk’s tunnel. They say they’ll stick with railways and roads." Virginia Mercury).

“It’s a car in a very small tunnel,” Michael McLaughlin, Virginia’s chief of rail transportation, told members of the Commonwealth Transportation Board’s public transit subcommittee...

But as a significant capacity enhancer, mostly Musk is talking about using these tunnels as fast tracks for single cars, or for small shuttles, such as for the project just approved in Las Vegas ("Elon Musk Company To Build Las Vegas People Mover," Engineering News Record).

That doesn't add very much capacity to the transportation system, especially with his idea for connecting the tunnels to the surface through elevators, which will significantly reduce throughput.