Driverless Cars — Taxi, Rent, or Own?

In my previous post on autonomous vehicles, I suggested that they probably wouldn’t come onto the market in the foreseeable future.  But I also suggested I might be wrong.  Let’s say I am wrong.

Let’s say that in the next ten years, you can buy a car that drives itself better than a typical human driver, under any “normal” street conditions (so, bad weather yes, traffic yes, residential streets and highway yes, but perhaps high-speed chase no or off-roading no).  And it is non-onerously legal to use this capability fully.  You don’t need a driver behind the wheel; you don’t need to put down a $100,000 bond for your robot.

What then?

Many people seem to believe that what happens then is that private ownership of cars essentially vanishes, and everyone takes (driverless) taxis.  But the people who think this seem an awful lot like they dislike American car culture and ownership anyway.  Is the idea that car ownership would disappear grounded in anything other than wishful thinking?

It’s grounded in at least one thing: we’d expect taxi rides to decrease pretty strongly in price, when you wouldn’t be paying a driver.  Let’s try to figure out roughly how much that takes out.

Here in foggy San Fracisco, the fare rules are $3.50 when you step into the cab, $2.75 per mile, and $0.55 per minute of traffic delay.  $0.55 per minute is $33 per hour, and if you’re going at 60mph on the freeway, $2.75 per minute is $165 per hour.  But of course cabbies don’t actually have their car full all of the time.  As it happens, I have access to data that suggests that cabs have a fare roughly 1/3rd of the time — so 20 minutes an hour.  Let’s call that an average of 2.5 trips per hour, and maybe an average of $50/hour for the time that people are in the car, plus a 20% tip.  This won’t be right, of course, but it’s in the right ballpark.  So, 1/3rd of $50 is $16.67, plus $8.75 for the beginning-of-ride bonuses, plus 15% tip, is just shy of $30 per hour.  Call it $30 per hour.  This handy report [PDF] suggests (on page 6) that it’s deeply unlikely that the median taxi driver takes home more than $20,000.  So of the about $60,000 that a full-time (8 hours a day five days a week) use of acab generates in a year, about $20,000 of it goes to the cabbie.

This suggests that very roughly 1/3rd of the cost of a cab ride is saved by doing away with the driver.  Now, that’s all back-of-the-envelope stuff, but the real take away is that taking away the cost of the driver doesn’t cause the cost of a taxi ride to just vanish.  You aren’t going to be talking about saving 90% of the present cost of the ride if the taxi is driverless.

Even if you get half off the present cost of a ride, taxis are still pretty expensive.  A ten minute trip at half present rates is going to set you back something like $6.  If you have a 30 minute commute each day, and do another hour of driving for non-work stuff, and you replaced everything with half-cost taxi rides, that’s something roughly like $216 per week, or $10,800 per year.  That doesn’t sound like a huge bargain compared to buying a car.  Even if my estimates are twice as high as they should be, $5,400 a year still doesn’t sound like a huge bargain compared to buying a car.

And let’s look at the ownership  model again, because driverless cars have big advantages for owners.  The most straightforward and probably the biggest one is that you, the owner, no longer have to put up with the hassle of parking and maintaining a drivable state.  If your car drives itself, you can have it drop you off just where you want, and then go park itself (or, worst-case scenario, circle the block endlessly if there’s no parking available), while you get your stuff done.  Your stuff may include drinking yourself well past the legal limit; no problem. You can nap in the car, or read.  These are huge advantages compared to the status quo!

Other advantages of the ownership model carry over.  A car is not just a transportation mechanism, it’s a mobile storage depot.  Think about how much stuff you leave in your car.  You won’t be able to do that with taxis, even robot taxis.  And many other advantages, besides price, contribute to the fact that the ownership model is the dominant mode of car transportation today.

Just based on that, I think that the smart money is on ownership remaining a fairly dominant model for transportation in the hypothetical robot-car future.  It’s dominant today; driverless cars make ownership more attractive; it may lose some market share, but lots of people will still own cars in the driverless future.

The other model that sometimes gets lost in the shuffle is longer-term rental.  Anywhere from ZipCar-like hourly rentals to traditional Budget-style daily rental.  The advantage that driverless cars bring to the rental market is that the lack of awkward arrangements to get to and from the rental location (always a pain in the neck when what you are trying to rent is a mode of transportation, but what you need to rent it is another mode of transportation, which is basically why rental cars have traditionally been an airport good), and that you can release the car when you don’t need it for a few hours.

ZipCar proves that there is at least reasonable demand for medium-term car rental, and driverless technology would allow them to serve larger areas without making hundreds of awkward deals for a parking space or two, as well as more effectively demand-balance.  We shouldn’t expect substantial reductions in cost, though.  ZipCars cost about $8 per hour for ordinary sedans in low-demand times.  Our six-hour a week driver (with a lot of his driving in high-demand times) is looking at a more reasonable, say, $70 per week, or $3,500 a year, which is starting to compare pretty favorably with ownership.  He also has the flexibility to rent a truck instead of his usual car, or a luxury car for date night, which is an advantage to him.

So I expect a driverless world to exhibit a heterogenous model of driving.  Still lots of ownership, especially for families with kids who want the minivan or SUV at least partially to haul around all their crap — putting child’s seats in a taxi does not appeal.  Lots of prestige car ownership.  The low end of the car market partially replaced by rental services and taxis. Probably modest bumps to the size of both the rental service and taxi markets.

But not the sea change of “hardly anyone owns a car.”

Disclosure: I work for Flywheel, which makes a taxi-hailing smartphone app.  The opinons expressed in this article are my own, and may not be shared by my employer.


4 thoughts on “Driverless Cars — Taxi, Rent, or Own?

  1. Personally I believe that when people speak of less people owning cars its not just about people using driverless vehicles as taxis. One option that was left out of your choices which I believe will be the broader model of car ownership in a driverless society is ‘car clubs’ where multiple people pay a small monthly or annual fee to have access/shares to a vehicle thus spreading the cost of ownership over more people.

    For one it is likely in my opinion that once cars can drive themselves that an entire population of people who currently do not have a driver’s license or who cannot drive will take advantage of access to transportation. These people along with those who currently drive will probably go out and buy one or two vehdicles instead of each and every person buying their own indiviudal car like most people do today. For example my wife and I work different schedules. I drive but she does not. Eventually she would want to buy her own car but in a driverless society she wouldn’t need to. I would get dropped off at work by my car and then it would go home until she neeeded to go to work.

    The reason we have so many cars is specifically because they can’t drive themselves.

    Most of the day and night a typical vehicle spends its time parked somewhere not doing what it is designed to do. Once cars are driverless I am certain that people will figure out ways to share that unused and wasted parked time with each other.

    Sharing this unused time would in my opinion drastically reduce demand overall for cars as people would have a cheap option for getting to where they need to go. All sorts of payment options could be made available depending on how sophisticated a car club you belong to is and how many cars are available to you.

    I would also posit that the main gist of the argument for driverless vehicles is one about safety and how much better they will drive than humans. The effect of all this safety will mean less wear and tear on vehicles, less crashes and thus less demand for new vehicles. Given a sophisticated vehicle that can drive itself and presumably monitor its own condition (and even likely drive itself to a mechanic when its feeling bad) you will certainly have a doubled effect of less demand for new cars and a sharp decrease in demand for cars overall as people begin to share their vehicle time with each other.

    In my last opinion, I believe the business model for auto manufacturs will be forced to change once driverless vehicles become ubiqitious. I just don’t see how they are going to continue to make as much profit selling cars to a society that won’t need as many (barring rises in population and as mentioned a wider variety of people able to own cars because they won’t need drivers licenses)

    I see a future where a successful auto manufacturer doesn’t stop at just making the vehicles but has taken on (or taken over) the task of creating customized ‘car clubs’ for people who want access to a wider variety of vehicles at better prices

  2. Hi Janarl,

    Thanks for the thoughtful comment.

    I would regard the car clubs that you envision as broadly similar to the ZipCar hourly rental model, at least once the car club crosses some threshold number of members which seems likely to me to be in the dozens. I agree that such a model is more attractive in the driverless car world.

    I wonder whether the demand for automobiles will actually decrease. Yes, I completely agree with you that some kind of demand aggregation will be facilitated (ie, your car clubs, or taxis, or rental), but also, as you point out, there will be increased demand from people who have traditionally been unable or unwilling to own an automobile (people who can’t or won’t learn to drive; people who are legally barred from driving (like, in a driverless world, there’s no real reason your 12-year-old can’t have his own car, if you’re wealthy enough to get one); people who find them inconvenient due to the difficulty of parking; etc.

    And while the number of accidents may decrease leading to less total replacement of vehicles, if the total number of miles traveled increases (because cars have more utility), it would be easy to see overall replacement over time go up, rather than down.

  3. Pingback: Who will own driverless cars? | Sandor at the Zoo

  4. Pingback: Uber’s Popularity Means Driverless Cars Won’t Be Ride-For-Hire | Sandor at the Zoo

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s