Forecasting is hard. If you think it is easy, find someone who predicted any three of these five things: that Chrysler would merge with Daimler, then demerge to be owned by private equity, then go bust, then be sold to Fiat, and then merge with Peugeot to become Stellantis. Right, didn’t think so. I certainly got no more than one of them right, at the time.
So, because it is so very hard, I usually avoid taking shots at automotive forecasters who make a blunder. Usually. Because in this one special case I am going to take the shot.
In 2017 I was working for a financial firm in Edinburgh (which gives you a hint as to my employer), when a US West Coast “think tank” called RethinkX released a report entitled “Rethinking Transportation 2020-2030”1. The authors said a lot of things about the future of transportation (we hadn’t relabeled it as the trendier “mobility” yet), and as a result I spent many hours and days patiently and not-so-patiently reviewing these things with my employer’s clients. It was infuriating, as many otherwise-sensible people were at that time drinking vast quantities of conceptual Kool-Aid, ingesting without question wild forecasts that I was tasked to review, discuss, and where necessary, refute. I took no rhetorical shots at RethinkX, limiting myself to dry discussions of facts and assumptions. (The late great Maryann Keller was not so reticent: on one call with her, her sarcastic disdain for the report was vivid enough that I thought she might melt a transatlantic phone cable.)
Well, the time has come to take a victory lap, because enough time has passed that we can actually test some of RethinkX’s 2017 predictions for accuracy. And I’ll just look at one - no need to pile on, right, especially at this time of year. So here’s the supporting chart, from the RethinkX report:
Yes, dear readers, the authors were so entranced by predictions of the imminent and widespread triumph of mobility services (aka ridehail, e.g., Uber and Lyft) and autonomous ridehail (aka robotaxis) that they predicted that in 2024 zero (nada, zilch, zippo) cars would be sold in the USA to individual persons for their own use. Instead of 17 million personal sales2 there would be just under 5 million institutional sales to “TaaS” (transportation-as-a-service) firms like Uber. Oh, and by the way, the report asserted these other developments as inevitably accompanying the utter demise of the personally-owned car:
“Car dealers cease to exist by 2024” - of course, yes, given zero retail sales
“New ICE vehicle sales are finished by 2024” - we’ll all ride only in EVs
And by 2030 (or earlier), traditional personally-owned ICE cars’ value will “plunge to zero or even negative value…. That is to say, owners may need to pay to dispose of their cars.” (emphasis added)
Again, readers, I spent a lot of time trying to explain to people, some of whom managed billions of dollars for investors, that these things just would not happen3.
So time for my year-end victory lap. I’ve stayed quiet throughout 2017, 18, 19, 20, 21, 22, and almost all of 2023, biding my time (revenge is a dish best served cold, etc.). but with about a week left in the year, and RethinkX’s 2024 forecast about to “go live,” I am here to tell you:
They were wrong.
I will go out on a limb myself now and predict: that at least a few dozen cars will be retailed in the USA in 2024, that at least a couple of them will have gasoline engines, that there will be at least three new-car dealers left in the USA, and that most of you won’t have to pay someone to drag away your model year 2020 Silverado.
My point? Beyond smug self-congratulation, it is this: the American automotive industry, and beyond that the transportation mobility behaviors of our 330 million fellow citizens (and their 285-plus million vehicles) are so broad and deep and vast and entrenched that changes to the ecosystem happen only slowly and at the margin, so we can actually see them coming. Our industry makes supertankers look lean and agile, and they take 2 to 5 miles to stop from full speed, and even their turning radius (to avoid obstacles) can be over a mile4.
Don’t get me wrong: I’m not against change generally5: to change is to be new, and to be new is to be young again, and we can all use more of that. And as for change specifically, I have nothing against robotaxis: look at the great progress Waymo is making. Maybe robotaxis eventually will be ubiquitous. After all, we went from elevators with operators to automated elevators, right?6 On the other hand, I am not getting on an airplane that has no human pilot any time soon. So I think of change, as this blog boringly reiterates, as a process of systems absorbing external shocks, responding to them, adapting to them, and eventually finding a new equilibrium. Forecasting is thus about finding that new balance, not just extrapolating the last three news stories we hear.
Thanks for reading this 10-part series. We’ll be back early in 2024 with a more measured pace of posts: I did this sprint mostly to see if I could pull it off.
No Dad Joke for this last post: the concept of Americans having to pay someone to haul away their used cars is funny enough in itself.
Here you go:
They got this wrong, too: some 15% of the 17 million, say 2.5 million, were indeed already wholesale units sold to fleets, like rental car companies such as Hertz.
At least to this extent and so quickly. Given enough years, or maybe decades, of course anything is possible!
See for example https://www.wikiwand.com/en/Seawise_Giant
Though frankly, I always thought that replacing Dick York with Dick Sargent as Darrin on Bewitched was the wrong move.
But I must point out that, according to BLS (Bureau of Labor Statistics) data, elevators still kill a couple of dozen people in America annually. Automating something does not make it inherently perfectly safe.