John Minchillo/AP Photo
Entrepreneur Andrew Yang, right, at the Democratic presidential primary debate in Westerville, Ohio, October 15, 2019
I attended a screening of Jojo Rabbit last night instead of watching the debate. I thought it’d be more entertaining and rational to watch the story of a boy whose imaginary friend is Adolf Hitler than CNN asking questions of presidential candidates. But I caught up later with a transcript.
Now, was that transcript generated by a person writing shorthand like a court reporter, or a machine listening to the audio? I don’t know, but the question became a topic on the stage last night, with a long section about automation and the relative merits of a universal basic income versus a federal job guarantee. And I think it badly missed the point, fundamentally misreading the current state of the economy.
Unsurprisingly, the exchange began with a badly flawed premise. Former Goldman Sachs financial analyst Erin Burnett cited “a recent study” claiming a quarter of American jobs could be lost to automation. The operative word is “could.” Predictions are hard, especially about the future. The estimates fall along such a wide range, between a trifling amount of job loss and half the workforce, that you cannot really say anything about what the future holds.
What we know about the last decade, definitionally the one with the most updated, cutting-edge robotics technology, is that the economy is behaving in the opposite manner as would one overwhelmed with automation. If you automate tasks in the economy, it should take fewer people to produce goods and services. The kiosks at the grocery store eliminate the need for clerks, so a store can be staffed with fewer employees. A robot-heavy assembly line doesn’t need as many steelworkers. That sort of thing.
But if you look at productivity statistics, you see that it has fallen, almost catastrophically, across the board in the last decade. There is no story about automation that’s consistent with output per hour falling precipitously. It’s literally the reverse of what you would expect.
Bernie Sanders kind of ignored that side of the debate by pointing out all the things that need to be done—repairing infrastructure, preparing to mitigate climate change—as solid reasons to guarantee jobs for those who want them. Sentient TED talk Andrew Yang rebutted this by saying people didn’t want to work for the federal government (highlighting his stay-at-home mom wife, when one of the professions Sanders noted could be part of a jobs guarantee was childcare), and later responded to Elizabeth Warren’s claim that job loss is more a function of bad trade policy by throwing out a bunch of anecdotes about retail stores closing and self-serve kiosks and self-driving trucks.
Again, this ignores the observable reality of the data showing productivity slowing down, not speeding up. Arguing from anecdote, and self-selecting anecdote at that (the kind of people that go to Andrew Yang rallies are obviously the kind of people attracted to his simple story about automation and a basic income), doesn’t effectively respond to what’s actually happening. Indeed, Yang citing his stay-at-home wife indicates how child care, for one, isn’t being pushed off to robots. Other high-cost sectors like health care, housing, and education have bumped along with similar manpower as well. The promise of massive online education courses has largely come to naught.
Here’s another example from one of the most high-tech companies in the world: Amazon. We know that the company is experimenting with cashless “Amazon Go” convenience stores, where you just pick up goods and have the price debited from your Prime account through your mobile phone. This is the area where automation is most acute, and because we see it directly, we think it’s occurring throughout the economy.
But on a much larger scale, Amazon is turning to an old standby to reduce labor costs: outsourcing. Hundreds of thousands of Amazon’s 2.5 million third-party sellers are suspended from the site or dealing with problems at any one time. The entire process for dealing with this has been outsourced to low-wage countries like India and Costa Rica.
Elsewhere, Amazon has tens of thousands of (outsourced independent contractor) drivers delivering packages to people’s homes and their car trunks, and moving to one-day shipping has put more of them on the road. This is slower-productivity manual work, relative to staffing an Amazon Go store where customers stock up. It’s just not consistent with a claim of widespread automation.
So why is overall productivity slowing? Is it trade policy, as Warren pointed out? Yes, outsourcing is a big part of the picture. But the real answer lies in something that Warren discussed later on in the debate: economic concentration.
If you are a dominant player in your market, you don’t need to invest in your business to maintain your share of customers. You “build a moat around your business,” as Warren Buffett is fond of saying, and that prevents competitors from crossing. This defeats the automation story by itself. Why would a dominant company invest in retooling equipment if it doesn’t need to? Especially as financial markets and executive compensation are guided by short-term profits, taking the capital expenditure hit to automate doesn’t seem necessary.
Indeed, the anomaly of the past few years, relatively unprecedented in economic history, is that investment remains relatively low while pre-tax profits have been at record highs. Incumbents don’t need to invest, and would-be competitors are afraid to, because they feel like it won’t be effective to dislodge the market leader. Investors are actually stumped about where to put their money, despite enormous social and economic problems throughout the country. It’s hard to get a return in a low-growth, low-investment, monopolistic economy.
This has created other ruptures in the social fabric as well, of course, in particular soaring inequality by class and by region, lower quality of service (why improve the product if it’s the only one on the shelf), and an economy dominated by financialization and short-term profits. This was also discussed last night, in a section that showed many candidates capable of tackling what’s truly holding down the economy. Sanders’s recent plan to establish market share thresholds for mergers, and Warren’s vows to break up large firms that distort markets, approaches the answer here. Public options and white-collar crime enforcement would also help. But it’s not a robot story; it’s an oligarch story.