Why Are Good Jobs Disappearing if Robots Aren’t Taking Them?
Post-Pandemic Automation Part I
By Data & Society Director of Research Sareeta Amrute, Sr. Researcher Alex Rosenblat, and Postdoctoral Scholar Brian Callaci
“The Robots are Ready as the COVID-19 Recession Spreads,” warns the blog of venerable think tank, the Brookings Institute. “Bosses Speed up Automation as Virus Keeps Workers Home,” The Guardian tells us. “Robots Welcome to Take Over, as Pandemic Accelerates Automation,” screams The New York Times. As of late May, at least 43 million people had applied for unemployment insurance after losing their jobs during the pandemic. Will these jobs be automated out of existence? Workers replaced by robots? Despite blaring headlines about self-driving trucks and self-checkout kiosks, the answer is almost certainly “no.”
…workers are regularly replaced by automation, but the speed and impact of robots taking jobs is routinely overstated in public discourse…
Though large-scale automation is not on the horizon in most workplaces, the nature of work is now different due to changes in the way work is organized, legal innovations, and shifts in the technologies deployed to oversee work in the last several decades. First, the nature of the workplace has shifted as companies have contracted work out to firms and individuals who do not have an employment relationship with the companies. Second, gig companies that innovate legally-contested and increasingly normative ways of managing work have changed labor and employment practices. Firms beyond the app-based gig economy are likely to take advantage of the legal claims gig companies have pressed to get around existing labor laws and worker protections. Third, surveillance, as well as predictive, and real-time just-in time scheduling tools have shifted the everyday experience of work for many workers, especially in service, warehouse, and care industries; under the pandemic, workplace monitoring for remote workers will likely increase. To be clear, workers are regularly replaced by automation, but the speed and impact of robots taking jobs is routinely overstated in public discourse; very unlikely to accelerate during a recession; and is one of many other factors that affect or change the nature of work.
Mainstream media describes automation with an uneasy ambivalence: a boon to minimizing human contact during the pandemic, and a menace to blue-collar jobs afterward. But the historical record paints a different picture. In the last three recessions since 1991, job growth lagged behind economic recovery, but not primarily because workers have been replaced by robots. Instead, firms restructure in response to downturns in ways that create fewer permanent job opportunities than in the past.
…companies use moments of temporary economic downturn to permanently restructure in ways that informalize work.
In other words, companies use moments of temporary economic downturn to permanently restructure in ways that informalize work. In the last three decades, companies have worked to “fissure” the workplace, offloading the costs of economic crisis by relying on heavily outsourced, just-in-time supply chains and contract workers. When the economies expand again, rather than hire more workers, firms increase their orders from outsourced suppliers and use temp agencies to fill shifts. When the next recession hits, they use these outsourced suppliers and temps as a buffer, cancelling their contracts, rather than making layoffs. The result is that “decent” work — a full-time, stable job with good pay and benefits — becomes increasingly rare.
The story of the fissured workplace begins with supply chains that, from the 1980s, distributed manufacturing and business processes across transnational boundaries. While new communications and logistic technologies aided these changes, they were just as equally the product of a shift in legal frames around work and a change in expectations about what kinds of work — and workers — need to be in-house, and what could be outsourced. Throughout the 1980s and 1990s, manufacturing and information processing jobs continued to be split apart, coordinated from the U.S. and Europe, but executed by vendor firms hiring local workers. These changes cheapened the cost of labor on the one hand and produced logistical supply chains on the other, with very little downtime or permanently-employed personnel. In India, the software technology parks and export processing zones that grew to meet demand came with legal exceptions to local employment laws and special rights to infrastructure that further enabled a shift in the workplace from in-house to dispersed production. These shifts in the legal, technological, and social organization of the workplace meant that work was spread across many nation-states, and that lead firms to displace responsibility for workplace protections onto subcontractors. These trends were put into place once again after the financial recession of 2008 by gig economy firms that practiced another kind of domestic outsourcing.
The so-called “gig economy” involves independent contractors completing tasks on-demand through an app rather than as employees of a firm with the rights and benefits that accompany employment. According to Silicon Valley gig companies such as Uber and Lyft, each ridehail driver is actually an independent entrepreneur running their own business rather than an employee, despite the fact that their drivers have neither the power nor information they need to make informed decisions. Drivers have little opportunity to take affirmative steps towards building their own business, such as maintaining client lists through the app or negotiating their own rates. Instead, they are responsible for running this “business” by maintaining the tools they need — like their cars — at their own expense, and responding to the incentives and penalties the companies use to manage them, even when these systems are opaque to the workers themselves.
Even before the recession, gig companies were leading the legal efforts to dissolve worker rights on several fronts.
In the U.S., electronically-mediated gig work only makes up about 1% of national employment, (though the gig economy is notoriously difficult to measure). But select gig companies are seeing a particular spike in visibility during pandemic isolation measures — like Instacart, which both became essential and saw huge, rapid growth.
Even before the recession, gig companies were leading the legal efforts to dissolve worker rights on several fronts. What is at risk now, is that the management techniques of gig companies will become protected by U.S. law and embedded in the national economy, while firms beyond the gig economy take advantage of the legal claims gig companies have pressed to get around existing labor laws and worker protections.
Given this convergence of factors, the important question is not whether robots will take our jobs, but how labor rights gain and lose authority alongside technological changes.
Zombie Robots: Coming, But Slowly
Despite years of hype, automation (defined as machines completing work tasks in greater numbers) has not happened. If automation were accelerating, we would expect to see a surge in labor productivity growth, since output per worker would necessarily increase. That has not happened. Rather, the rate of productivity growth has been slowing, not accelerating, in recent years. Since 2008, labor productivity growth has been almost as anemic as it has been at any point since the Great Depression. It is still far from its postwar peak of the 1960s.
At the same time, investments in capital equipment, information processing equipment, and software — all things we would expect to precede and accompany automation — have been slowing down since 2000, not speeding up. And these investments are set to decelerate rather than accelerate during economic recessions. If the future is anything like the past, the pace of automation will continue to be mild, and will only pick up (moderately) during the recovery to come.
This is the first blog post in a 2-part series. If robots aren’t about to take all our jobs during the time of COVID-19, what is going to change? And how is new technology likely to be used in the workplace? Read about this in part II.
Sareeta Amrute is an anthropologist and author of the award winning book, Encoding Race, Encoding Class: Indian IT Workers in Berlin.
Alex Rosenblat is an ethnographer and author of Uberland: How Algorithms are Changing the Rules of Work.
Brian Callaci is an economist and has recently published in Phenomenal World and The American Prospect.
