Self-Driving Cars or Robot Taxis?
A Revolutionary Choice
The most significant development in self-driving cars so far this year launched on March 21 in Altamonte Springs, Florida. Working with Uber, this suburban city of 42,000, ten miles north of Orlando, has agreed to a year-long pilot project during which the city will pay up to 25 percent of the cost of Uber trips within its borders.
Altamonte Springs and Uber are directly confronting an unavoidable truth about how Americans have built cities and suburbs. And that truth is not about cities and too much parking.
To gauge whether transit network companies — with their ride-hailing apps, car rentals metered to the minute, and promises of self-driving cars — are indeed disrupting personal travel, keep a close watch on what happens to vehicle ownership.
According to the U.S. Census Bureau, over 90% of American households have access to at least one vehicle — a car, van, or small truck — for personal use. Yet it is commonplace for transportation pundits to assert that most people won’t own self-driving cars. Their line of argument boils down to something like this: why go through the hassle and expense of owning a car, when you can summon one at will?
Uber and Lyft, and their high profile dives into robot vehicles, have added momentum to the presumption against ownership. Media coverage of the race between Google and Uber now refers not to cars, but “self-driving taxis.” And during the public-relations blitz following GM’s recent investment in Lyft, corporate officials trumpeted a future of networked rides, not buying vehicles.
By 2040, assuming government really does ban human-driven vehicles as some experts predict, if ownership falls from today’s 90 percent of households to around 80 percent, the impact will be significant, but not revolutionary.
If instead vehicle ownership, nationally, falls to 70 percent, 60 percent, 50 percent or even lower, the disruption will force the massive reshaping of American cities, suburbs, and rural areas. Every ten percentage points would mean another 20 million households adopting car-free lifestyles by 2040.
The American automobile society built during the 20th century was fundamentally about owning cars. For the vast majority of Americans, personal travel has meant the option to hop in a car whenever you want, to go wherever you want. But owning a car does something else that 100 years ago was revolutionary in personal travel: it lowered the marginal cost of the next trip to just pennies per mile.
Since the 1920s, American cities, suburbs, and rural areas have been built around the perceived price of driving the next trip, not the total cost of driving. For most car owners, the perceived price of the next trip is just the cost of gasoline.
At the dawn of the motor age, cars did not simply replace the horse and carriage in the city. According to historians Clay McShane and Joel A. Tarr, only the top of the urban elite could afford to keep horses in cities, mostly for recreation and — quite literally — for publicly parading their wealth. This was because keeping a horse in a city cost hundreds, if not thousands, of dollars a year.
American automobile manufacturers — most famously led by Henry Ford — designed their products for the average American family. By the early 1920s, a new Ford Model T could be had for less than $300. By the end of the decade, historian Kenneth Jackson estimates one car for every five Americans. That works out to 24.6 million cars for 123.2 million people.
Today in the United States, total spending related to driving ranges between $3,500-$4,500 per vehicle per year for low- and middle-income groups, according to the Bureau of Labor Statistics’ most recent figures from 2014. For the highest income groups, spending jumps to between $5,000-$6,000 per vehicle per year.
Spending on gasoline, however, is remarkably consistent across all income groups at about $1,300 per vehicle. The cost of gasoline ranges between 23 and 37 percent of total driving-related expenditures. Roughly two-thirds (or more) of driving-related spending is, for all practical purposes, a fixed cost.
In other words, roughly two-thirds of total spending on driving is the price of admission to a car-owning lifestyle. This lifestyle is about more than speed and convenience. It is also about the consumption of space as measured in miles of road, acres of land, and square feet of housing. And one man’s urban sprawl is another woman’s room to breath.
For the person who hires trips in taxis, however, 100 percent of the cost of driving becomes part of the price of admission to the next destination. To this or that restaurant. To this or that medical complex. To the community college in the next town or the low-wage job in this town.
The economic consequences of the difference between the price of admission to a lifestyle and the price of admission to the next destination is why ownership matters.
But without drivers, won’t rides in robot taxis be so cheap that most Americans will be able to afford them? Thanks to the IRS, it is fairly easy to conduct a thought experiment that explores travel decisions in a world dominated by robot taxis that by some miracle cost the same as driving today.
Every year the IRS announces a per-mile reimbursement rate for driving personal vehicles for business purposes. It covers fixed and variable operating costs and includes depreciation, insurance, repairs, maintenance, gas, and oil. The IRS rate estimates the total cost of the next mile driven.
What the IRS rate doesn’t include is compensation for the driver’s labor.
For 2016, the IRS reimbursement rate is 54 cents per mile, down from 57.5 cents in 2015 and 56 cents in 2014. To reimburse yourself for driving 7,500 miles per year equals $4,050, comfortably within the average spending per vehicle.
If you drive more in the 11,000-mile range, the average miles traveled per vehicle in the United States, the IRS rate now costs significantly more per vehicle — $5,940 — than the typical American spends on driving. Depreciation bites.
What does this look like per trip? The perceived marginal “just gas” price obviously varies with fuel efficiency and gas prices, but for the purposes here I’m using my most recent figures: 10 cents per mile. The IRS rate is 5.4 times higher than my “just gas” price.
Driving to the pool to swim laps, a four-mile round trip, costs 40 cents at the “just gas” price. At the 2016 IRS rate, to go swimming costs me $2.16. If I actually hit my goal of swimming 12 times a month, the cost is $25.92, as compared to spending $4.80 for “just gas.”
But the real question is what does paying the IRS rate per mile feel like? Try it yourself. Imagine paying 54 cents per mile for your daily trips. Imagine choosing between restaurants, deciding whether to meet friends, or deliberating which medical specialist to consult. Imagine it is the end of the month and you are trying to stretch every dollar. What happens?
Behavioral economists call this emotional experience a disruption in mental accounting. Under normal circumstances most people do not comprehensively re-evaluate all of their spending whenever prices change. Instead, they adjust something “nearby.”
For transportation, mental accounting means that it is a mistake to assume that former car owners will simply make a dollar for dollar substitution from their total budget for ownership and driving into their budget for taking robot-taxi trips. For one thing, the shift moves transportation spending from an annual or monthly budget of mostly fixed costs into the daily or weekly budget for variable spending.
In addition, Americans are sensitive to small changes in the perceived price of the next trip.
Take gas prices: true to mental accounting, the first way Americans respond to increases in gas prices is to shop for cheaper gas. Americans may explore other transportation options like walking, bicycling, and taking transit, but for many, after cheaper gas, the next option is to drive less. Keeping close to home — or even staying home — is a transportation choice.
Consumer responses to recent drops in the price of gasoline confirm that “nearby” does indeed mean other daily, variable expenses. Americans have spent 80 percent of their savings from gasoline, mostly on eating at restaurants and shopping, according to a recent large-scale analysis by JPMorgan Chase Institute.
For a robot-taxi world to feel like today, fares would have to approach the “just gas” price, not the IRS rate. Another way to think about it: if the IRS rate were my “just gas” price, gasoline would cost over $10 per gallon.
(Charges for parking have the potential to close or reverse the gap between the prices of admission to the next destination experienced by owning a car versus taking trips by taxi; but the car owner would have to pay for parking hourly or daily, everywhere and always.)
And here is the Catch-22: for robot-taxi companies to get the volume of ridership needed to miraculously lower the price to even just the IRS rate, they need people to give up vehicle ownership. But for people to give up ownership, the price of the robot-taxi trip needs to come way down. When people choose car-free lifestyles, however, they have a huge incentive to live in places that have transportation options besides taking a trip in a robot taxi.
Where are places that today support car-free living? Where vehicle ownership rates are below 80 percent of households? At around 70 percent are the cities (not including their suburbs) of Chicago, Baltimore, Philadelphia, and San Francisco. Coming in at the 60 percent range: Boston, Washington, DC and Queens, NY. To get below 50 percent requires moving to Brooklyn or the Bronx. Manhattan is in the low 20s.
Together, these cities contain only about five percent of American households.
And so it makes sense that suburbs and Uber are already looking at public subsidies. In Altamonte Springs, where today 96 percent of households have access to a car, the pilot project supports an additional transportation option. In a world of robot taxis, continuing subsidies from suburbs may very well be a matter of their survival.
We’ve had fun spinning fantasies of parking lots transformed into parks. But it is time to start asking who lets go of ownership and how this changes their decisions about where to live, work, shop, and recreate in a world where robot taxis charge — with every trip — the total cost of driving as the price of admission to the next destination.