Reading the back-and-forth, the takeaway for me was just how much pure guesswork - on both sides - went into the estimates. It's not just the professor's arguments that are easy to poke holes in, it's every single valuation you can think of for Uber. Here are three scenarios, for example, where Uber value trends toward zero:
Google brings self-driving cars to market. Instead of owning a car, a fleet of autonomous vehicles drive around and pick up any passenger who requests it. No drivers are necessary, the app is built into the Android OS, and all the logistics are handled by Google Maps.
Workable telepresence solutions are developed, relying on combinations of e-mail, web applications, GitHub, videoconference, and some yet-to-be-invented virtual presence system (holodecks? Oculus Rift?). Instead of requiring that you be in a physical office, all knowledge workers can telecommute from wherever they live, and their "office" exists only in virtual reality. Seeking cheaper real estate, larger houses, and yards, workers leave cities in droves. Socializing happens in virtual worlds, and all those trips that Ubers are currently necessary for never happen.
The taxi union lobbies hard and gets Uber and similar ride-sharing services declared illegal in all major cities.
And that's discounting the truly black-swan events like "World War 3 breaks out, and we're all wiped out by nuclear winter."
It really drives home how the market value of a company is set by the last person to buy a share of that company, no matter how crazy or irrational his beliefs are. I could momentarily drop the market cap of Google to $6.75M by selling a share at $0.01, but presumably the market would correct itself in a millisecond. In thinly-traded private stocks, this market correction mechanism is not always available, and so valuations can be way off the mark until all the assumptions built into those valuations have become public and been judged as likely or not likely by the market.
Not read the Tesla piece but at least with the Uber piece the assumptions were clearly stated so they can be challenged.
One challenge I would make to this article's assumptions is that the world is like San Francisco. In many places there are already cheap private hire options available[0]. I'm not sure how to calculate for the existing cost and availability of such options. If it makes sense for the drivers they may become Uber drivers but there may be less decrease in car ownership and changes in customer behaviour.
To add to that, what about those cities where you have also a good public transport in place like NYC, Lodon, Paris, Berlin and many more? Where it is may be comfortable to use such a service, but also more expansive and more slower (because of traffic jam).
London not only have good public transport and horribly slow/congested roads in the centre. We also have cheap minicabs, Hailo, and at least one "Uber-like" that contracts minicab companies rather than individual drivers. And many of the larger minicab fleets seems fairly similar to Uber in the way they operate as well.
(We also have rickshaws; mostly as a tourist novelty, though)
Doesn't mean they can't do well here too, but certainly a different proposition to one where they only competes against normal taxis.
Yeah, but tesla as a car company is overvalued right now and was then, and last year that's all it was or looked to ever be.
This article is refuting the market for the product; I don't think anyone believes the market for electric vehicles is significantly larger than the current automotive market.
The issue with using a DCF based approach, is that you have to bet only a single scenario whereas if you take a decision tree/probability based model, you can account for a range of scenarios. Here is Charlie Munger on Warren Buffet:
One of the advantages of a fellow like Buffett, whom I've worked with all these years, is that he automatically thinks in terms of decision trees and the elementary math of permutations and combinations....
====
Here is a probability model vs DCF based model. Goldman uses a probability based model which can be used to evaluate changing shifts in technology such as Tesla's Autonomous technology
Google brings self-driving cars to market. Instead of owning a car, a fleet of autonomous vehicles drive around and pick up any passenger who requests it. No drivers are necessary, the app is built into the Android OS, and all the logistics are handled by Google Maps.
Workable telepresence solutions are developed, relying on combinations of e-mail, web applications, GitHub, videoconference, and some yet-to-be-invented virtual presence system (holodecks? Oculus Rift?). Instead of requiring that you be in a physical office, all knowledge workers can telecommute from wherever they live, and their "office" exists only in virtual reality. Seeking cheaper real estate, larger houses, and yards, workers leave cities in droves. Socializing happens in virtual worlds, and all those trips that Ubers are currently necessary for never happen.
The taxi union lobbies hard and gets Uber and similar ride-sharing services declared illegal in all major cities.
And that's discounting the truly black-swan events like "World War 3 breaks out, and we're all wiped out by nuclear winter."
It really drives home how the market value of a company is set by the last person to buy a share of that company, no matter how crazy or irrational his beliefs are. I could momentarily drop the market cap of Google to $6.75M by selling a share at $0.01, but presumably the market would correct itself in a millisecond. In thinly-traded private stocks, this market correction mechanism is not always available, and so valuations can be way off the mark until all the assumptions built into those valuations have become public and been judged as likely or not likely by the market.