Do the economics of self-driving taxis make sense? | Hacker News Reader

Do the economics of self-driving taxis make sense?

ftalphaville.ft.com

37 pointsjoosters56 comments

The main point (and fairly interesting critique) of the article;

    Except there’s a major problem with that utopian
    view. The vision totally underestimates the true value
    humans bring to the car market or the degree to
    which Uber’s “urban saturation” depends on a
    dedicated human element to absorb the bulk of the
    capital cost and risk of public vehicle ownership, with
    very little personal reward.

    As Felix Oberholzer-Gee, a professor at Harvard, 
    tells Nautilus, if Uber can’t outsource costs and
    market fluctuation risk to contractors, especially the
    cost of maintaining idle fleets during off-peak hours
    directly to Uber, it quickly swaps an asset-light
    business model for an asset-heavy low-margin
    model instead.

    This is problematic because Uber’s already low-
    margin business model depends heavily on being
    able to overstep regulation, licensing and — most
    importantly — transfer maintenance, cleaning,
    insurance and market-risk exposure to drivers. It
    also benefits from drivers’ ability to draw value from
    vehicles in personal time.

The article sorta veers around so it's hard to isolate its main arguments. I'm not sure why Uber is really the relevant model however. A more obvious starting point is autonomous taxis. And then the question becomes: Are self-driving vehicles really all that more interesting because you eliminate a $10/hr driver? For a lot of scenarios, I suspect the answer is not really.

According to http://www.taxifarefinder.com/main.php?city=SF&from=San+Fran..., 34 min taxi ride will cost you $104.95.

An hour of driving a taxi would cost you at least $150.

There's maybe $10 of non-human cost (car usage deprecation, fuel, insurance).

ZipCar, for example, is less than $10/hr and cheap car rental is $40/day.

Self-driving taxis, due to higher utilization and less management over-head, should be cheaper than car rental and more convenient for people.

So we're talking about at least 10x rate reduction for auto-taxis than current taxis which will change behavior in a big way (lower prices will lead to higher usage which leads to higher profits).

Ultimately the economics of self-driving taxi-like service should be: it's cheaper than total cost of ownership of owning your own car, even if you use it frequently (~2hrs a day).

I like the comps you use - we have baseline companies providing just part of the equation, like ZipCar, so we just have to extrapolate for the tech overhead to auto-drive.

You bring up a valid point, but it's worth considering that Taxi drivers have homes. Let's think about driving from SF to Palo Alto at 11pm. This is probably not great for a taxi driver who: a) might have wanted this to be his last shift b) lives closer to SF

There are other ways of selecting for this, but a self-driving taxi doesn't need to care and can just be managed as a fleet. Yes, some miles will need to be driven by an empty car, but for the most part demand for transportation has a balance, but this balance is very hard to reach in practice with actual drivers.

There are certainly scenarios where the cost delta between our hypothetical fully-autonomous vehicle and a taxi as we know it today is greater and where it is less.

I'd argue that in Manhattan, the distances are short enough and the utilization high enough that self-driving probably buys you very little over a taxi today. With the scenario you describe, the cost difference is probably greater. (On the other hand, if that's 50% of today's taxi costs, most people won't be using it to commute every day.)

Of course, there are still other cases where it's not feasible for most people to hire a driver today--say going away for the weekend because the driver costs start to dominate the vehicle rental costs. (And in that scenario, you just rent a car today if you don't own one.)

In any case, when such vehicles become available in however many years, they will probably reduce car ownership. But, as with Zipcar, I suspect the difference will be at the margins.

I actually think NY is a pretty compelling case for driverless. A large number of people go from hubs into the city, travel around the city, and then leave the city.

Also there's an issue with balancing taxi load in the outer boroughs vs midtown / downtown manhattan. There's a reason uber has been so successful even in NY at increasing the total number of passenger trips and not just stealing customers away from traditional taxis.

A piece discussed here last week (a FiveThirtyEight blog) seemed to suggest that Uber pickups in NY have mostly come at the expense of taxis rather than being accretive.[1] Though, anecdotally, Uber works significantly better for the outer boroughs than taxis do.

NYC experiences are hard to generalize though. And traffic probably means you can't inject many more vehicles around peak times. My basic point was that, given high utilizations and somewhat balanced traffic patterns, taxi/Uber drivers don't add a huge amount of cost relative to the vehicle itself.

[1] https://news.ycombinator.com/item?id=10387310

I see where the author is coming from, but uber doesn't need to own the autonomous vehicles. They can still be owned by third parties who maintain them. The owner just "recalls" the vehicle when he needs it,

Exactly. Anyone who owns an autonomous vehicle will be able to offer it to Uber for their use. The only difference is that they won't need to also drive the vehicle.

The point about outsourcing vehicle ownership to the drivers is a valid criticism of Uber, but self-driving cars doesn't really need to play into that equation.

With any luck, there will be more than one marketplace for buying and selling vehicle miles.

As a former part time uber driver, the challenge is convincing people to send your car out to uber alone. Even with me in the car, at a minimum passengers would leave trash or gum.

>at a minimum passengers would leave trash or gum.

Doesn't the review system handle this? Goes back to one of the other points the article makes: this system depends upon your personal information being available to uber as well as the ability to 'blacklist' bad users.

one complication to that is groups of passengers, where only one person's info is available

Funnily enough, this comes back to "ride sharing" which Uber etc. supposedly were in the first place. I'm skeptical that this would become a thing--or at least mainstream. But I suspect that we're talking enough years out that so many things will be different that predictions are tough.

Yes, the option of having your autonomous vehicle go off and work as an uber driver while you're at work is a point in favor of Uber continuing as a broker, and keeping out of the fleet management/capital investment business.

Which makes it confusing that Uber has made noises about wanting to buy self-driving Teslas in the past...

Exactly. There will be sales pitches when fully autonomous cars are available to buy: pay $xk up front, but lease the car to a fleet for $y/month to recoup some of your investment. Uber might require owners to maintain and clean the cars in order to qualify, saving themselves the cost.

It's not dissimilar to the way people throw their houses into the pool of accommodation that is Airbnb or VRBO, etc.

>They can still be owned by third parties who maintain them.

Exactly. See: Hertz's partnership with Lyft. They've anticipated this already. Whenever the car is not needed, it's sitting on a Hertz lot ready to be rented as a private vehicle.

    Except there’s a major problem with that utopian
    view. The vision totally underestimates the true value
    humans bring to the car market or the degree to
    which Uber’s “urban saturation” depends on a
    dedicated human element to absorb the bulk of the
    capital cost and risk of public vehicle ownership, with
    very little personal reward.

Self-driving taxis won't compete with privately-owned cars. They will compete with buses.

There will be a strong case for municipal transit authorities to invest in them instead of buses, light rail, and other services that cost more and do a worse job of taking people where they want to go.

It sure makes sense, but unlike private newcomers like Uber, those are all unionized shops and unions will fight hard to block any such improvements because they would mean heavy job losses and unions exist to protect jobs.

See BART strikes.

Most likely they'll achieve pyrrhic victory: muni will be it's usual self for a couple of years, private system will reach cost parity but will offer better service, bus ridership will plummet, eventually it'll become evident that it doesn't make sense to pump more public money into it and it'll either be disbanded or possibly replaced by a contract with private company (because it'll be hard to rebuild muni to be competitive price-wise after giving private companies years of head start).

A contrarian view: asset-heavy business is good for the winner because of barriers to entry. How many competitors SpaceX will ever have?

So if Uber wins the market, they'll have it for themselves.

Asset-heavy transportation businesses aren't as risky as people assume. Airlines buy they airplanes. Shipping companies buy their (massive) ships. Hertz can afford the cars they rent. FedEx can afford their truck.

Those businesses rarely go out of business and the "market fluctuation" is not really a thing. If there's a sudden 10% drop in flights, the airline will bump the prices by 10%, retire some routes and voila, they are profitable again.

I get that there's capital and risk benefit to offloading car ownership, I don't see how that translates to low margin.

Margin is whatever Uber decides it to be. If they can lower the cost of providing the service by 10x by not using humans, then they can offer a price that is only 5x lower and get obscene 50%+ margin.

What might affect a margin is lots of competition, but this seems like "winner takes most" business (currently Uber is 10x Lyft, it's closest competition in US).

Finally, getting money to fund such business is easy. Uber already raised ~$10 billion. Assuming cost of self-driving car to be ~$20k, that's half a million cars.

Given current market condition (low interest rates, lot's of money floating around), a business with a plan will have no problem raising essentially unlimited funds.

The main point is missing the forest for the trees. The main benefit of self-driving cars is improving efficiency and productivity. It is productivity that makes us all much richer.

Which business model self-driving taxi operators go by is pretty irrelevant; the free market will settle on the best way. The productivity increases will benefits the entire society should be the most important part of any analysis.

But the article is about self-driving taxis. There is little productivity gain there for the users, since they are already free to spend their time in the car working or doing whatever they like, driver or no driver. Likewise, self-driving cars won't make everyone richer, taxi drivers will definitely be on the losing side.

> It is productivity that makes us all much richer.

It is productivity that makes the owners of the productive assets richer, not "us all" (whether that somehow reaches others is dependent on factors other than the productivity boost itself.)

Nope even if you don't own assets you still benefit from increased productivity. The average person don't own airliners, but still benefit from cheap air fare tickets.

With a free market and self-driving taxis, the price of transportation will collapse which will benefit everyone. The wealth that used to be spent on transportation can be used to by other goods, which will make everyone richer.

This is incorrect. In order for the asset owner to gain from an increase in productivity they to interact with customers, workers and suppliers who in turn interact with others. This ends up sharing much if not most or even all of the gain with others. You are correct that many factors will determine how the gain gets allocated but only in exceptional circumstances does the owner keep all the gain.

https://www.boundless.com/economics/textbooks/boundless-econ...

When you factor in the safety features of self-driving cars, the automated traffic management/reduction in the number of private cars, and the ability of less-well-off people to get to work in less time and make it home to see their families... seems to benefit everyone.

If we can reduce the cost of transportation (even through subsidies) for the poorest, that opens up opportunities.

Uber doesn't need to own the cars to provide self-driving service. Suppose I as a consumer can purchase a self driving Tesla that is fully autonomous, and can have instructions provided remotely. I could use some kind of Uber app to have the car go ferry around passengers when I don't need it. Depending on how much I use the car and demand for rides, I could actually turn a profit on the month's car payments. Uber maintains it's decentralized network and I absorb the costs and risks of vehicle ownership presumably with some kind of self driving taxi car insurance. Many more people would likely join the uber network once they can do it without actually having to drive.

The problem with asset-heavy models isn't the cost of capital, which is extremely cheap these days. The problem, as the article alludes to, is that regulator or unions might shut it down, and then their expensive asset sits idle.

That only happens if you have a single market without production mobility between them. But if one city shut down Uber's self-driving cars, they could quickly truck them to a new city.

Perhaps my perspective is too nescient (I'm an engineer — not an economist), but this claim seems sketchy to me:

> How does [sharing] add value to the economy over the longer term? There is no production. It is entirely consumption. Recycling is all very well – and we do need secondary markets – but we cannot build an economy solely on sweating existing assets. An economy that exists solely on consumption has no long-term future.

Sharing is a process that optimizes usage of available resources. Why should we make more cars than we need? Why should everyone have a boat if not everyone needs a boat? Sure, more boat-building jobs are "created" if we feel everyone really needs their own personal boat, but there seems to be a strong opportunity cost that is being forgotten here: the labor necessary to build boats could instead be directed into a more useful form that advances society as a whole much more quickly.

What do I mean by advance? Well, I guess to some degree that's subjective. But I'm fairly certain everyone can agree on at least some universal goals that humanity would like to achieve, and artificially increasing the number of jobs available for the sake of having more jobs available seems like it would hurt our overall progress in the long run.

I think this quote from Warren Buffet sums up what I'm trying to say much better than I can:

> The way I see it is that my money represents an enormous number of claim checks on society. It is like I have these little pieces of paper that I can turn into consumption. If I wanted to, I could hire 10,000 people to do nothing but paint my picture every day for the rest of my life. And the GNP would go up. But the utility of the product would be zilch, and I would be keeping those 10,000 people from doing AIDS research, or teaching, or nursing. I don't do that though. I don't use very many of those claim checks. There's nothing material I want very much. And I'm going to give virtually all of those claim checks to charity when my wife and I die.

A old Canadian critique of jobs for the sake of jobs:

> Taking up the policy of a public works program as a solution for unemployment, it was criticized as a plan that took no account of the part that machinery played in modern construction, with a road-making machine instanced as an example. He saw, said Mr. Aberhart, work in progress at an airport and was told that the men were given picks and shovels in order to lengthen the work, to which he replied why not give them spoons and forks instead of picks and shovels if the object was to lengthen out the task.

That reminds me of the hypothetical idea of central banks increasing productivity/employment by burying money in the ground. Suddenly you create lots of mining jobs and shovel sales. Or the 'broken window' fallacy.

> Sharing is a process that optimizes usage of available resources. Why should we make more cars than we need?

Isn't it weird how we spend so much time talking about conservation of resources and then an article like this will make the same people say, "Sharing leads to no production!". Like you mentioned, this is simply a matter of humans getting better at making things and using the things we make. We'll find new things to make.

Production != progress. Increases in wealth == progress.

I don't think the article is trying to imply that we should be making stuff just for the sake of making stuff, or creating jobs just for the sake of creating jobs. It's just giving an explanation for why many new industries don't show up in productivity measurements.

If car sharing did take off in a big way, then it will redistribute value. Presumably fewer cars will be sold, so car manufacturers will be worse off. The people sharing cars might be better off (they would earn money through sharing, but note that they are sacrificing several things for the added income), and the companies running the sharing business definitely become wealthier.

That Buffet quote certainly reminds me that the rich are different from you and I. He could also hire those 10,000 people to do AIDS research, etc., but he really wants to die with the most toys. "Charity" can have them afterwards.

That's certainly true for many super-affluent people, but Warren Buffett is hardly the poster child. He lives a relatively modest lifestyle, and has already given away a large portion of his fortune.

I agree, he seems not to have lost his moral compass -- he's no Larry Ellison. But something about the phrase "enormous number of claim checks on society" suggests that he's collecting points, and trying to rack up a high score by the end of the game.

I'd argue the super-wealthy are only different than the "rest of us" in terms of scale on this issue.

Relevant Louis CK:

https://www.youtube.com/watch?v=lC4FnfNKwUo

I drive a cheap Honda Fit, not an Infinity, but the concept still applies to almost everyone who lives a decently comfortable western life.

If Buffett wants to try to rack up a high score, and plays by fair and equal rules in the process (not hurting anyone, not pressuring people into transactions, paying his taxes on time, et cetera), what's wrong with him doing that?

I'm veering off-topic here, but Buffett's large deals are often cunningly crafted so that taxes are eliminated. He's got the right to do that, but it doesn't make him a shining example of a taxpayer.

Still, the charities he is giving his wealth to will get some of this benefit!

To me, it suggests exactly the opposite. Rather than simply seeing wealth as a high-score indicator, he recognizes that by holding onto it he is squandering a huge amount of potential wealth for society at large.

One huge benefit to Warren Buffett waiting until he dies to give his money away is that he has consistently achieved outstanding returns on his capital, so by holding onto it now, he has much more to donate tomorrow. If he had given away all his money when he had $1 billion he wouldn't have $50 billion to donate today.

Interestingly enough, all of the bullet-pointed problems are problems inherent in physical plant management for manufacturing (except privacy). This is possibly due to the lessons capital-intensive industries that have to date been manufacturing are also applicable to capital-intensive services that are largely commoditized and interchangeable (like getting transported from point A to point B). Taxi service is a good textbook example of a service that requires a minimum level of knowledge, but is extremely repetitive and doesn't benefit exponentially from human intervention.

Since they didn't provide any numbers, here's an analysis I did on this topic using estimates for LA: https://www.linkedin.com/pulse/how-cheap-can-uber-self-drivi...

Uber can absolutely make money even if they own their own cars, provided the up-front cost of the cars is low enough. You can think of it like a high-use, low-distance rental car fleet - the same car economics down-sides apply to Hertz and Enterprise and yet they still make money.

And as self-driving fleets take over and cease to be tied in to the driver's self-image, my guess is that the cars will look more like the $15,000 Toyota Yaris. I also imagine that as fleet owners gain greater control of design, they'll find ways to make maintenance easier and cheaper. Insurance should also get cheaper with fewer human drivers (and by the time automated cars are on the road, there will be a lot of cars with crash-mitigating technologies too).

This article and the 2 linked source articles are a big hot mess of half baked thoughts and misunderstandings. However it did remind me again to think of all the impacts that driverless cars will have.

Does anyone know of a thoughtful analysis of how things (like public transit, parking, etc) will likely change as driverless vehicles become more common?

Brad Templeton's essays are the place to start: http://www.templetons.com/brad/robocars/

IHS Automotive forecasts that the price for the self-driving technology will add between $7,000 and $10,000 to a car's sticker price in 2025

Uber will probably use it's own fleet supplementing people during peak times. if you can get 70% utilization that's about 60k/yr you don't pay a driver.

Do you happen to have a link to this forecast? I've never heard of IHS Automotive but it sounds interested. Could the 7-10k increase be offset by other efficiencies in engineering that bring down the cost?

Side note, could we be entering the era where one person does not own one vehicle, but many people own a few? I could see something like an apartment complex offering a few self-driving cars as an amenity, or cities investing in them like they do bike-shares. Lots of the discussion has focussed on the extremes: individual ownership or megacorp ownership. There might be a place for this technology in between.

Not sure which IHS article OP is referring to, but BCG did an analysis as well: https://www.bcgperspectives.com/content/articles/automotive-...

Relevant section is "The Price Tag". Here's a helpful image from the paper as well: https://www.bcgperspectives.com/Images/Revolution_Drivers_Se...

The biggest cost driver is the LIDAR, which is anticipated to cost Google $8k per car next year when they move to a cheaper system.

Anyone else getting a paywall?

FT has a free registration as well, do it once, stay logged in.. Their content is pretty good all around.

I was a long time subscriber to FT and still think it is above average but sloppy articles like this make me think the quality has been declining with the rest of the industry.

view the site via google (search for the title, click on the result)

While I appreciate the help, I shouldn't have to.

This has been discussed before. Please see [0].

[0] - https://news.ycombinator.com/item?id=10178989

Reply on news.ycombinator.com