31 August, 2015

Uber (Part 4a) - Consensus view of on-demand ride economics and the implications

Part of my beliefs on Uber is that marginal cost pricing will drive the end-state economics of both drivers and platforms (4.b and 4.g in the framework).  Because of this, I'll (later) share specifics of how I believe vehicle cost curves are relevant to Uber's strategic imperatives.

My original plan to bring readers up to speed with the "consensus view" of expert analysis in this corner of the business world was to post my own narrative over the weekend - including some wonky discussion of a recent NBER working paper titled "Peer-to-peer markets," authored by Stanford's Jon Levin et al.

Then I read Alex Rubalcava's piece "A roadmap for a world without drivers," published last Wednesday on Medium.  Alex is a VC in a "family office" firm with several investments in the automotive space.

Rather than reinvent the wheel, I offer you Alex's good-enough summary consensus (slightly reordered, in bold, with my beliefs indented below).

TL;DR All you really need to know is that experts agree this industry is headed towards marginal cost pricing, with or without autonomy.  Autonomy could help Uber in the medium-term, if it can survive that long

Alex's autonomous vehicle discussion is essentially the same economic structure as our Uber discussion, except with driver costs approaching zero.  For further detail, read his piece and in particular follow his links to expert published views on various aspects of the discussion.

Primary (to our discussion) summary points: (from Alex's page 2)

  • Without the need to pay drivers, Uber-like services with autonomous vehicles (AVs) will cost 50% to 90% less than they do today.
    • Agreed.  The portion accruing to the vehicle (with/without driver) will approach variable cost of operation as explained in the paper by Prof. Levin
  • Uber, or someone like them, will offer autonomous shared vehicle services. Widespread adoption of such services would reduce the amount of vehicles needed to accomplish the same amount of transportation by between 75% and 90%, depending on assumptions about utilization rates and consumer preferences for sharing rides.
    • I'm not convinced that... 
      • These services would be offered by Uber
      • If they were that Uber would make nearly enough money from offering them
        • To justify their current valuation
        • To afford the capital to make the transition (see the framework in my second Uber post)
    • I also believe autonomous vehicles won't reduce the number of autos in use by nearly that amount except in the densest of city cores.  More detail on why I believe this in my upcoming Uber (Part 4c)
      • Also, if the number of vehicles were reduced significantly, car prices would skyrocket as costs are kept low through broad amortization of design costs and high-fixed-cost production.  Like other industries, the remaining auto manufacturers are already struggling to balance product breadth with sufficient production to remain cost competitive (and Sergio Marchionne keeps getting rebuffed by Mary Barra to discuss this)
      • Therefore if autonomous does actually mean fewer overall vehicles, it's more likely that auto manufacturers will get into Uber's business than Uber will get into theirs.  This is why you see auto manufacturers doing autonomous vehicle research at Tesla, Audi, BMW, even Tier 1 suppliers like Delphi.  
        • Baidu (Chinese version of Google) claims to be releasing an autonomous BMW in China later this year
        • Uber board member Steve Jurvetson is claiming Uber CEO Kalanick has already offered to buy all 500,000 autonomous cars Tesla plans to make in 2020.  
          • Electric is definitely the way to go.  More in Uber (Part 4b).
          • Would that be better for Tesla than selling cars-as-a-service on their own?  I suspect Tesla could recreate Uber much easier than the other way 'round
  • Much of the urban real estate dedicated to parking will get re-purposed
    • Agreed.  And few (if any) suburban parking areas will be eliminated
  • Vehicle miles traveled per person will not change much from today
    • Agreed, particularly because of "empty" re-positioning trips

Secondary (to our discussion) summary points: (from Alex's page 2)

  • There will be much less traffic, even if shared AVs do not take off, from the more efficient driving patterns of AVs
    • Not necessarily.  This assumes that the number of vehicles in use drops dramatically.  Today, upwards of 30% of the peak-hour traffic in urban cores such as NYC has been shown by multiple studies to be cars circling the block looking for parking.  
      • If parking is eliminated and the number of vehicles in use doesn't drop significantly, it's possible that non-shared autonomous vehicles would circle the block while waiting for their owners
      • Or at least drive to outlying "holding pens" until they are recalled, like when family comes to pick you up at the airport
  • The improved safety of AVs will reduce insurance premiums between 50% and 90%
    • Probably, although this is an unhelpfully wide range.  Liability should decrease remarkably and transfer from operator to manufacturer.  General collision would presumably decrease as well.  Rates are already shifting to mileage-based, as that's how risk accumulates.  This point is basically saying a non-fuel variable cost will be lower, more accurately priced, and transparent.... and I agree
  • The first vehicles on the market will arrive between 2017 and 2020, with rapid changes to the transportation infrastructure following in short order.
    • I'm skeptical of the speed and adoption, but am not close enough to have an informed opinion.  Alex believes it will take longer.  I tend to agree, though not necessarily with his reasoning

VMT rebound as discussed by Alex is probably overstated

Alex also makes his own prediction of what he calls a "Vehicle Miles Travelled (VMT) Rebound."  His belief is that because rides will get much cheaper, we will consume much more of them - especially because passengers can work in cars if they don't have to drive them.  This view is naive for two reasons: low price sensitivity of demand, and almost universal incidence of motion sickness for passengers of autonomous cars not paying close attention to the road.

The limits of a VMT increase by current drivers are likely much lower than Alex believes.  Gasoline prices have little impact on VMT today; current estimates are a 25-50% decrease in fuel costs create a 1% increase in demand.

An electric-powered autonomous vehicle cost per mile, even at variable cost pricing, is unlikely to fall much below $.20/mile (more soon in Part 4b).  For comparison, fuel costs of a 30 mpg car when gas is $2.50/gal is $.083.

Even once accounting for maintenance and depreciation, the variable cost for driving a Chevy Spark today is about $.20/mile; a Honda civic is closer to $.25 - this is on par with the variable cost of a fully autonomous, electric-powered fleet of on-demand cars.

Any surge in VMT would likely be due to eliminating externality costs (such as parking downtown) or avoidance of mass transit for people who currently can't afford the fixed costs of car ownership.  The first one is likely relatively low.  The second one could be particularly pernicious, as even a small initial erosion could have a cascading feedback effect on the viability of public transit in the long run

My own belief is that a fleet of on-demand autonomous vehicles would mostly put more money in peoples' pockets - much like lower food costs have done in past decades.  Perhaps some will splurge with their newly found wealth by travelling in a BMW instead of a Civic.

As an aside, as a passenger in a car I'm unable to physically do anything except take a phone call or watch the scenery.  And lest you think that I'm mistaking myself for the market, studies are starting to find that motion sickness in autonomous vehicles are likely to be much worse than in regular cars.

So a dramatic VMT increase from the existing automotive driver/passenger base caused by (a) lower prices and (b) the ability to conduct other work as a passenger doesn't seem very likely to me.

Thanks for reading,

No comments:

Post a Comment

Focused, topical critiques are welcome though my responses may be limited. Off-topic comments, trolling, spam, or anything not pulling its weight will be deleted.