Has Uber already beaten private ownership on cost?


Today, various experts, like CR and the AAA rate the cost of private car ownership anywhere from 40 to 60 cents per mile, plus parking. That depends on your usage patterns, what car you buy and its age, plus a few other factors. Many people, though, pretend that using their car only costs the 8-12 cents/mile for gasoline. (A better estimate of the truly incremental cost without factoring in those things that don't vary with the miles is around 25 cents/mile.)

Riding in an Uber costs from $1.50 to $2/mile, and nobody has any illusions about it, because you pay the full cost right there. At 3x to 4x the price (or 7x the price for those who only count the incremental costs) it's a luxury.

I have often reported here on how robotaxis will cut the price to 25 to 40 cents/mile blowing up the economics of car travel. That will happen, but what's surprising is it's already happening.

Today, many people are switching to services like Uber over use of their private car. That's because they are factoring in the value of their own time, as well as the cost of parking. I have been doing this myself, and there's an analysis of the costs and the switching at TheConversation.

Even though I own a car, it's been a couple of years since I drove it to the airport. The combination of the cost of airport parking, the cost of my own time, and the convenience and time-saving of going right to the curb rather than taking a shuttle from a parking lot have made it a win. The 35 mile trip tends to cost about $55 (x2) in Lyft/Uber and that's a lot more than the $35 to use my own car at 50 cents/mile. Add just a few days parking, though, and it starts to get close enough that it's easy to make up the rest of the difference from the value of my own time.

I charge my clients good money for my time, and while most people aren't quite willing to value their own time at the rate they sell it for, every middle-class and above person should value the 1.5 hours involved in driving and shuttles at at least $30.

Today, I just fiddle with my phone on the trip. In the future I'll be able to do a lot more. The fake "robocar" is already winning even at 3x the price.

Other analysis has shown that in the dense cities which have very expensive parking (and personal garages) that using ride services is a win even without putting more than a minimum wage value on your time. For example, parking at San Francisco hotels is $35 to $50 per day, making it easy to justify Uber use for a 2-3 conference where I stay in the hotel.

The first robotaxis won't be 50 cents/mile; I think that $1/mile is a likely early target. Even that should blow the doors off private car use, at least financially, for many people. At 50 cents/mile (or my predicted 30 cents) the private car is going to have a very hard slog.

Not all private car ownership will go away. Outside cities, robotaxi service will not be conveniently available on short notice. Some people will just resist the idea. Some people love to keep stuff in their cars and have it available everywhere they go -- particularly families with small children. But for the rest, the economic pressures will be very strong.

Cars designed for the rear seat

One thing I expect we'll see soon is a car from the major automakers designed primarily to be used with Uber and Lyft.

Of course, there have always been high-end luxury cars -- limousines -- aimed at the chauffeured driving market. And there are cars like the Nissan NV200 and the London Taxi Company TX4 designed for that market. The London vehicle, with its rear facing seats, is the most different from a typical vehicle.

A car sold to be an Uber/Lyft would still be somebody's car, unless it's one of the ones leased just for this use. But many people almost never use their back seats -- or even passenger seats -- and so they should be open to more radical designs.

I would like to see:

  • The ability to recline, or even sleep.
  • Power ports for charging my devices
  • A large screen I can connect to for video or working
  • Tray table (I've seen that in a few offerings)

Today, because back seats are mostly unoccupied except in family cars, not much attention is given to them. They usually don't have any recline (and can't because of the trunk behind them.)

One design might be to have the rear seat on the passenger side flip up so that the front passenger seat can go back 1 or two feet, halfway into the back. Then it could recline, and get a privacy shield. (You can't put in a privacy shield now because the driver needs to see out the passenger side.) That's good for one, but doesn't handle two.

Naturally such a car would command a higher rate, though if it gets too much higher, it breaks some of the economics I list above, where driving and being driven are near equal in cost.

How much cheaper will private car ownership get?

Private car ownership will also drop in the robocar world. An electric sedan, with lower insurance costs, minimal parking costs, lower fuel costs and lower interior costs (no need for adjustable everything or fancy infotainment) will be cheaper than today's cars, though not as cheap as the same thing half-size for one passenger.

The other key advantage of the taxi is it wears out in 5 years while the private car takes 20 years. When a car is a computer, who wants a 10 year old computer? In fact, the private robocars will need to be designed so their computing and sensors can be field upgraded if things are going to work at all.


Take 100 or any number of families that want to own a robocar. They make a club, as any club of today, but instead of play golf or tennis, the members can use the "premises" (robocars) of the club for free or a fee. According to the schedule, the robocars that are free for a period, can be hired for other users, and the club get the fee. If a member needs a car and no one is available can hired from another club or for Uber, and the club will pay that. Call the club "UBER", and the members will have the services than uber can give, etc, but the profit will go to pay the maintenance of the club. All the expenses to buy the car, insurance, maintenance , others, are pay for the club, and the members pay their memebership. The type of cars and amount of each type, will be decided for the members. Let it club make his own rules.

Yes, I used to write about this idea more back in the prior decades, but the big question is, what's the virtue of the club? You don't get to keep your stuff in the car and have it 100% for your use as is the case for private ownership. So what do you get? Mainly, for some more money, you can give your fleet lower utilization so that perhaps club members can get shorter wait times than the public does from the big services. But why can't the big services just offer that as a premium feature, and do better than the club? The big service will offer access to classes of car type, or luxury levels. So what does the club offer that justifies its higher prices?

To talk about a RC club , is just to explain in a simple way what would be just a pool of RC, like an UBER, that owns all RC in an area. About private ownership:

1) The ownership of a RC will be different than a car (you wrote about this). The owner will be not be responsible, for flat tires, bad battery, bad brake conditions and others; the RC will take care of himself. In the meantime another RC from the club will be available

2) the use of a RC for a family will be different than the use of a family car today. Each member of a family from 2 to 120 years old can use a RC independent of other members of the family.

3) But today Daddy left home 2 hours before he needs to do so, to bring mummy to the mall before 1 hour it opens, and then the child at school 40 min before school begins, then drive to work and left the car 8 hours paying parking fee, at 1km from office to pay less money. Everybody wastes time.

4) A family owned a RC will do near the same if there is no time to go and return to pick up each member of the family. If the RC can go and return, will be more costly in running, and yes, will save parking fee waiting at home. Why not be used for another members of the club, instead to be parking?

5) Owning a RC, will not avoid the need of hire a Uber RC sometimes, why not a RC of my club?.

6) Owning a car today or a RC in the future, to keep staff inside, is not useful when a member of the family needs a taxi today or a RC Uber tomorrow, cannot have the staff that is in the family car at that moment. So really we need the staff ALWAYS in the family car?

7) A RC risks to become obsolete very fast, loosing his value. So the most hours a RC is used will be better. A family RC cannot be used 24 Hs a day, a RC from a club has more chances to do so.

8) If a member wants or needs to own a car for 1 hour, 1 day, 1 month, 5 years, to keep staff inside, or any reason, can do. Just pay the tariff for that period of ownership.

9 ) Any club can buy different types of cars, so members can use the appropriate car each time.

Please put blank lines between your paragraphs!

1) May be in the xx century, more people have died or seriously injured in car accidents than in all wars, massacres and terrorist attacks. If not, anyway the numbers are huge.

2) Only few countries took the things seriously. With 1,200,000 deaths from car accidents per year worldwide, means that not too much have been done.

3) Robocars can take those numbers near zero, but depends in which way we introduce them, the variation could take many, many years, too many.

4) After a short test period that already started , and may be will last less than 2 years, a worldwide RC deployment must be done.

5 ) Reasons why RC and man driving cars MDC (man driving cars) cannot coexist.

A) Suppose 20% or less RC in an area. Car accidents will be not less than 20%, not too much, but ALWAYS a MDC will hits a RC or causes a RC hits another car or a person.

B) People will start to be reluctant to take a MDC instead a RC, because always MDC makes accidents and rarely RC.

C) Driving tests will be more difficult and more drivers will be banned for driving because their bad behavior while driving (passing a red light, alcohol, speed, etc), not just a fine.

D) RC cannot be made less stronger (and lighter, saving energy), because the risk to be beaten by a MDC.

E) Services for MDC (mechanics, gas stations, parking, others) ,will start to diminished, causing problems.

F) Same for RC, services will grow slow, because still not enough RC.

G) All things will cause a fast grow in the desire to have enough RC in the area, causing social unrest.

6) So we must decided if we want RC to avoid fatalities and besides have a better transportation system an other advantages or will let the things running according to the profit desires of car manufacturers and other companies. They can take profit anyway, but let us organize the things for better.

7) In any area, once RC begin to grow over some number, owners of actual cars will find real big difficult to find services for their cars, causing social unrest plus RC shortage in the area, plus more unrest.
BRAD, you wrote that at the beginning traffic will be worse. I agree. With a fast deployment of ALL RC in an area, will be no problem.

H) how to deploy and organize RC is for another post.

Virtually every calculation I have seen comparing robocars or robotaxis or whatever to privately-owned cars assume that the relevant privately-owned car to compare the robotaxis to is a NEW car. Does this make sense, given that a) we buy some 40 mm used cars each year, versus some 17 mm new; b) the average age of a new-car buyer is easily over 50? The typical American in their 20s and 30s is buying a used car, at more like a $15,000 average price versus a $30,000 or so new-car price. By using new-car prices as the benchmark, are we making an error? Is this like comparing renting a house to buying a new house, given that "used" houses dominate the real estate market? I don't have a strong view on this, just wondering what are the right numbers to use.

This is a fair comment. In theory, robotaxis will never be very old in years, because they should wear out in under 5 years if they are engineered for the 200,000 mile lifetime of today's cars. They could be engineered for a longer lifetime, which could make them a bit cheaper, but it does mean you have them out on the roads 9 years old with 400,000 miles on them, and this will not be appealing to customers unless you are refreshing the interior on a regular basis, and probably refreshing the digital hardware which will be obsolete before 9 years.

Uber is so sure that people don't want an old car coming to pick them up that they forbid vehicles over 15 years.

Of course, for the first decade there will be no old robocars avilable to own. Purchased robocars for the first few years will be high-end vehicles as well.

My forecast calls for new robocars for 1-2 people to cost less to buy and operate than an older used car. To buy because they are smaller and do not have all the complex features needed for drivers. To operate because as newer vehicles, they will have lower repair and maintenance costs.

Add new comment