The peculiar economics of autonomous cabs 

Ashley Nunes is a research fellow at Harvard Law School. His work explores how innovation affects markets.

Lyft riders are in for a treat. Some of them anyway. The ride-hailing giant recently announced that riders on five per cent of trips will soon be ferried around by robots, not human drivers. Autonomous vehicles are a “huge opportunity”, said Lyft president John Zimmer, who plans a hybrid network that will “scale up with our autonomous partners.” Those partners include industry heavyweights like Ford, Argo, Motional, and Waymo, all of whom are itching for a piece of a robocab pie.

That pie is purportedly the stuff of dreams. Investors — SoftBank, T Rowe Price, and Microsoft to name a few — may have spent millions on bringing robocabs to market, but the prize is worth far more. Intel-owned developer Mobileye pegs the total addressable market at $160bn, a conservative figure by the company’s own account. Chinese investment group CITIC Securities says the true figure is closer to $500bn. And never to be outdone, ARK Invest claims Tesla’s robotaxi platform alone could be worth upwards of $9tn over the next decade. If you are looking for a place to park some cash, robotaxis are, it would seem, a safe bet.

Such hefty optimism reflects how market players think about so-called mobility on demand, aka the taxi industry. A sizeable chunk of taxi revenue goes towards paying drivers. Cut out the driver and a financial windfall awaits. Eradicating human labour costs is particularly attractive to ride-hailing companies keen to balance the books; organisations whose financial statements have been described as “a haemorrhaging fountain of red ink with no path to profitability”. Driverless tech should change all this by axing out their largest operating expense.

Such reasoning makes intuitive sense. But intuition isn’t always right. And where robocabs are concerned, intuition is dead wrong.

For one thing, driver salaries, although significant, affect revenue generation far less than the taxi’s utilisation rate — the per cent of miles it travels with a fare paying passenger in the back. And that’s a problem because in some cities, drivers spend a fraction of their time earning revenue (the rest is spent finding it). This supply-demand matching inefficiency isn’t restricted to any one market. From Singapore to Shanghai to San Francisco, cabbies struggle to consistently keep the back seat full. Robocab technology isn’t designed to address this problem. The result? Poor utilisation — and by consequence lower revenue — regardless of whether humans or machines are at the helm.

A larger problem is the myth of full autonomy. When technology is perfected, human labour demand will — we are repeatedly told — be eliminated. Yet, autonomous does not mean humanless. In Our Robots, Ourselves: Robotics and the Myths of Autonomy, MIT historian David Mindell explains why.

“There are no fully autonomous systems,” Mindell argues. “The machine that operates entirely independently of human direction is a useless machine. Only a rock is truly autonomous.” Put another way, the type of automation the likes of Lyft are betting on to boost balance sheets does not exist. It never has, and it never will.

Which probably explains why Lyft’s Zimmer envisions deployed autonomous cabs on just five per cent of trips. That’s a far cry from 2016 when his predictions were more — to put it mildly — ambitious. Back then, Zimmer released a 14-page manifesto in which he said that by 2021, “autonomous vehicle fleets will quickly become widespread and will account for the majority of Lyft rides.” He also predicted that robocabs could provide, compared with personal car ownership, “better service at a lower cost.” It turns out that’s not quite true either.

Resisting the financial allure of autonomy isn’t easy. Why resist an easy windfall made possible by algorithmic prowess? This explains why Wall Street, Silicon Valley, and government officials for the matter continually tout the virtues of autonomy. But autonomy never seems to quite arrive.

Maybe it’s time we accept that. Maybe it’s time we acknowledge that where AVs are concerned, investors would benefit from more intellect and less temperament. Maybe it’s time we see autonomy for what it is: a Disneyland-style spectacle that can’t, by one account, “live up to its sci-fi imaginings, a series of very expensive and glitzy pilot projects that can’t cut it in the real world.”