More on Uber “Competing” with Transit

On the whole question of whether ridesourcing (Uber/Lyft/etc) can replace fixed route buses, my fullest explanation is here, in both text and video form.  Since then, there’s been some interesting news:

  • Uber lost over $1.2 billon in 6 months.  Yes, with a “b.”   Cite this next time someone goes on about “money-losing” public transit systems.
  • Uber is starting to do absurdly deep discounts that look a lot like predatory pricing.  20 rides for $20????.  Such a price would undercut transit, thereby causing massively increased traffic with all the resulting ills.  It’s obviously unsustainable for Uber, but if it goes on long enough to damage transit systems, that will be a huge negative impact on our cities.  I hope someone is looking at whether this is legal.
  • In happier news, California has its first example of Uber/Lyft/etc. replacing a fixed bus route, and it’s a very sensible one.  It’s at the Livermore / Amador Valley Transit Authority on the eastern edge of the Bay Area.   The key is that the bus route being replaced has predictably dreadful ridership: only five people get on for every hour that the bus operates.  (Usually, a poor suburban fixed route performance is at least 10 boardings/service hour.)  This is almost as bad as a short-trip taxi, which means that Uber/Lyft, with the ability to pick up multiple riders on the same trip, might do just as well.  It’s still not a clean replacement.  The fare is higher than the transit agency’s, as it must be to compensate for ridesharing’s inefficiency, and transfers to the buses don’t appear to be free.  This is a reasonable deal, but there are  not many fixed bus lines that perform this poorly!


13 Responses to More on Uber “Competing” with Transit

  1. Alan August 29, 2016 at 5:38 am #

    From the linked article, it sounds like the $20 fee is not the fare for the rides themselves, but is the cost to “reserve” a flat fee as the fare for 20 rides. The fares themselves will be $2 for Pool, and $7 for X, so the total cost per ride will be $3-8 – and only for trips that begin and end in the designated zone. Even the Pool fare would still be higher than transit, at least in SF – whether that would be the case in other municipalities they might try is an open question.

  2. Mike August 29, 2016 at 5:11 pm #

    Except for being operated by Uber, this kind of service is not new at all, and has been used in Canadian cities for decades under the name Taxi Bus or TranCab.

    They are often used to connect neighborhoods or industrial areas which are beyond the reach of a regular service route. The service generally takes riders to the nearest regular service bus stop.

    Transport Canada has a mini paper on these services, and how cost effective they are, with some services recovering upwards of 45 to 50% of costs from the fare.

    The one issue I see with this Livermore / Amador Valley Transit Authority example, is that the neighborhoods losing fixed route service would have regular service if they were in a Canadian city. Planners have to stop using “suburbia” as an excuse for poor transit service. Maybe the 5 boardings an hour was because the service was poor?

    • Dave August 30, 2016 at 6:55 am #

      While I largely agree with your comment, there are two things that set Canadian cities apart from American ones that cause American cities to generally not have transit service coverage or frequency in suburban areas:

      1. The cost (in money and time) of owning a car in Canada is higher than in America. Gas in Canada is far more expensive, as is insurance… and most Canadian cities and towns (even the small ones) seem to charge for or heavily regulate public parking spaces, unlike in America. Furthermore, Canada really didn’t go on an urban highway building binge in the postwar era like America did – for example, Vancouver doesn’t have a massive highway slicing its way into downtown like Seattle does. This all means that suburban Canadians are more economically incentivized to try riding transit than suburban Americans are… and once commuting patterns are established, they tend to stick.

      2. Canadian urban areas did not experience racialized class conflict the way that American urban areas did. Downtowns in Canada weren’t completely abandoned by anyone who had money in the period of 1960-2000 like they were in America, which hollowed out core cities (and their transit ridership) that could have been used to cross-subsidize new service in the suburbs. Of course, the few Americans left behind in core cities kept riding the bus, but that gave transit the appearance of being a “poor” and/or “non-white” thing to those viewing it from their new homes in the suburbs. And since those new suburbanites feared having to move yet again if “those people” they left behind in the city moved out to the suburbs too, they actively fought any attempts to extend transit beyond the areas where it already existed prior to WWII.

      Unfortunately, the combo of both these factors results in American metropolitan regions having far less transit service than they “should” when compared to any other comparatively rich nation. And with its usefulness being both highly dependent on surrounding land uses and also highly contingent on network effects of having other lines to connect with – transit is one of those things that once you kill it, it’s hard to resuscitate it. American cities will probably always be playing catch up to Canadian/European/Asian cities in terms of public transportation.

  3. asdf2 August 29, 2016 at 10:12 pm #

    I suspect part of Uber’s business model is that the fares of people who use up all 20 of their discounted rides will be subsidized by those who don’t. 20 rides in one month is a lot, especially considering that the service area for these discounted fares is the section of the middle of the city that has the worst traffic and best transit service. It’s easy to see people signing up for the service imagining they’ll use their 20 discounted rides, but ending up not actually doing so.

    The other part may be an effort to try to improve the productivity of UberPool. The few times I’ve ridden it, I’ve had the car to myself, and I was basically getting UberX for about 20% off. Perhaps with this “flat rate” model, many of the extra trips that taken will end up filling empty seats in UberPool trips that are operating anyway, so the additional miles traveled, may be less than it might seem.

    Finally, this service could be a loss leader to simply get people in the habit of using Uber for not just oddball trips, but everyday commuting, in the hopes that it could eventually be replaced with something less costly to operate (for example, a service where the driver is commuting to work, just like the passengers, and gets compensated only for gas).

    Of course, one thing missing from this whole discussion is how the drivers of these “flat fare” trips end up being compensated. Do they still get paid for the miles and minutes and surge, as if the customer were paying the regular metered rate? Or, does Uber just keep the subscription revenue for itself, and give the driver just 80% of the already-discounted price?

  4. P August 30, 2016 at 11:46 am #

    Uber has lost a lot of money, but there is one difference between that and public transport services. Investors who voluntarily put their money into that business are who wears the losses or gains, whereas with a public service, its voters who have to wear the losses.

    It is doubtful hat Uber will damage urban transit systems. Transit systems are government subsidised and don’t need passengers to run (see all the empty coverage services that run with low or no patronage). Often a transit agency will have their passengers signed up to long-term transit passes, so it will be a sunk cost if pax move over to Uber to test it out. In the long run, what is not sustainable will not be sustained.

    If anything, Uber might push more cities to look and review their transit operations, which may initiate long needed bus network renewal. More business for transit agency consultants I suppose!

    • Steve C August 31, 2016 at 4:04 am #

      Yeah. Transit’s always just a boondoggle. Not like highways. A new highway is always justified and a wise expenditure. We should all just abandon cities and worship the big box with “Acres of Parking!!!”. The best part is that it’s all on the government’s dime. It’s free. We don’t have to pay for it. Wow! It’s like magic.

      • Transit Riding Transit Planner September 2, 2016 at 2:12 am #

        I think this reply kind of misses the point of P’s comment. The simple fact is that public transit systems are better equipped to operate at a loss over a long period of time than private companies like Uber. In fact, all North American public transit agency budgets and operating plans explicitly require them to do so! And no, I’m not talking about the so-called “coverage” routes in the ridership/coverage dichotomy frequently discussed on this blog. The vast majority of “ridership” routes also have a less-then-100% farebox recovery ratio. Observe that the busiest transit system in North America – the NYC subway system – has a 50% farebox recovery ratio. The highest farebox ratio I’ve heard of for an urban transit system is BART’s at 67%. You could probably cherry pick especially busy lines or sections of such systems and calculate them to be operating at a profit, but you’d have to disregard network effects in such calculations.

        This is in no way a dig on transit, and really has nothing to do with highways. The failures of privatized transit have been well documented here and elsewhere. For a variety of reasons, publicly funding transit has turned out to be the preferable model – in fact really the only model that has worked at all on the scale of an urban network for the past three quarters of a century. This, in itself, is neither good nor bad, but it does happen to be a huge structural advantage if a private company decides to start a price war with you.

        • ararar September 2, 2016 at 4:28 am #

          when it’s a city service that the tax payers pay to provide to themselves it’s hard to make direct comparison to private businesses who only have to look at their bottomline imho.
          The NYC subway system allows a level of mobility that allowed the economy to flourish.
          So looking only at the farebox is reductionist imho, as the return on investment can also come from the bigger tax base generated in the limited space that is NYC.

  5. R. W. Rynerson August 30, 2016 at 11:01 pm #

    It’s interesting to see the rate of ten passengers an hour, as in my rules of thumb for service planning, that’s the dividing point where customers believe a service is necessary. Below that, they get nervous on a regular route. On a call-n-Ride at five passengers an hour they see it as popular.

    In Colorado there’s an interesting experiment started by the city of Centennial with grant funds to use Lyft for “last mile” service radiating from the Dry Creek RTD LRT station. Media people are excited that it will run free (till the money runs out) — but do not mention that it has to run free because it is running head on against an established RTD call-n-Ride that has free transfers with the regional network.

    Free travel will give it an advantage for trips internal to the call-n-Ride zone and Lyft offers a shorter lead time for booking trips. Reading between the lines, Lyft is expected to have a lower operating cost. Though the call-n-Ride is run by a contractor, its operators are expected to be professionals, paid accordingly.

    For info on the call-n-Ride:

    For info the Lyft offer:

    In peak hours the call-n-Ride offers a semi-fixed route which is driven by the fact that the same people book rides every day. The free Lyft service could attract some riders who would be on the outer end of the “route”.

    It’ll be interesting to see if the added service cannibalizes the market or expands it. And, of course, when the grant runs out, whether the usual practice of trying to get RTD to pick up the tab will be followed.

  6. Jay September 1, 2016 at 2:52 pm #

    Uber’s fares are remarkably cheap. But I’m not sure how long that can last. I think that Uber’s business model relies on the drivers not realizing the true cost of driving their own cars. Sure, they are making $15 or $20 per hour but they are basically destroying their car in the process, especially in tough urban diving conditions.

  7. Dexter Wong September 2, 2016 at 12:26 am #

    The Verge has published an article on a small Florida town that subsidized Uber to provide transport to residents who had smart phones or credit cards.

  8. EHS September 5, 2016 at 12:34 pm #

    I was thinking about this as I rode the bus the other day, and I came up with a new argument that I’m finding pretty compelling. Wanted to run it by the wise folks here to see what y’all think.

    Question: Why do people ride a bus that doesn’t have exclusive right of way? It’s guaranteed to be slower than driving on the same streets.
    Answer: It’s cheaper in costs other than time. The fare is cheaper than driving, there are no parking costs, you can use another mode for your next leg without stranding your car somewhere, you can do something while you ride, your capital costs are a bus pass and not a car loan.

    Question 2: Why do people sit in traffic next to transit that’s going faster than they are? (Look at any freeway-aligned rail project during rush hour for an obvious example, though the alignment, and argument, are much stronger if not along a freeway)
    Answer: It isn’t easy to get out of the car and onto the train/BRT. You’d have to find a place to park. If you could park, you wouldn’t have the convenience of a car when you get off the train or bus, so you might have a long walk, or a slow transfer to a slow bus to get to your destination. You’d have to come back to where you parked your car so you could take it home.

    OK, so let’s look at what happens when cars don’t have drivers, shall we? Well, first thing is that the cost of uber/taxi options will plummet, because there isn’t a driver to pay. It should approach the cost of running a car, with just a little margin to sweeten the deal for the car owner and to pay for somebody to connect cars with passengers efficiently (that margin might even be smaller than the interest payments on a car loan, in which case people who couldn’t pay cash would chose to use driverless). The more people do this, the more likely it is that there’s an available car nearby when you want it, and even one with a passenger going the same direction. Suddenly, you’ve got efficient, point to point carpooling. With a few people in a car, the costs are significantly below those of running your own car.

    OK! So we’ve got a brave new world where there are cars zooming around every which way (or ahem, stuck in traffic) without drivers, and you can hop in one to go where you want to go, pretty much at will. There will of course be traffic jams of empties going reverse-peak to pick up their next ride, but so long as driving empty costs less than the money made from the peak direction passengers and the savings on parking, traffic jams of empties will be a thing.

    Now, how does this interact with the scenarios in the two questions above? Well, the driverless car looks a whole lot better than the existing car that’s competing with a slow bus. You don’t have to buy it, you don’t have to park it, you can do whatever you want while you’re in it, and, if you’re willing to carpool, you could save a good hunk of change, too. If it’s an improvement on an existing car, it will be better competition with a slow bus – this doesn’t bode well for the bus. Reduced ridership means increased fares or subsidies, and probably reduced frequency, which is terrible for ridership. Ugly.

    But what about the second scenario? A fast train fares far, far better. You can just have the driverless car drop you at the next station, and, bye bye traffic! No worries about parking. If you want a car when you get off the train, you just pick one up. As long as you aren’t transporting significant cargo, you’d be a fool to ever take a car along a route where a train is faster. (It’s even possible that driverless cars could set up transit to impose minimum speed for traffic along a corridor. Other than local traffic and cargo, it wouldn’t make sense to drive a route much slower than transit.)

    The bottom line is that the effect of driverless cars is dependent on the quality of the transit affected. Fast transit will get a boost, slow transit will be hurt. To Silicon Valley types that are loathe to plan for a future governed by rules of today, because tech will revolutionize the future: get hopping on grade-separated transit.

    (notes: This is a user-choice perspective, rather than the geometry of cities perspective that planners like Jarrett Walker tend to think in. I’d argue that geometry is how you figure out what should happen, user choice is how you figure out what’s most likely to happen. Also, there are huge questions about how driverless cars will effect land use, which has bigger effects on transit than driverless cars could ever hope to have directly.)

  9. Dexter Wong September 12, 2016 at 11:39 pm #

    This just in from The Verge: Ford buys Chariot: