Covid-19 will only increase automation anxiety   

This article is a part of a series in which the Financial Times asks leading commentators and policymakers what to expect from a post-Covid-19 future 

The writer, author of ‘The Technology Trap’, is the Oxford Martin Citi Fellow and directs the Future of Work programme at the Oxford Martin School 

Historically, depressions have inflamed automation anxiety. Wars, in contrast, often end them. In the US, the machinery debates of the 1920s and 30s began with labour secretary James J. Davis’s famous 1927 speech. Yet panic only set in during the Depression, and only faded with US entry into the second world war. As Covid-19 is often described as a warlike event, and one that has also tipped the world into recession, will it end or exacerbate our automation concerns?

One reason the second world war brought automation anxiety to an end is that everyone had to work at full capacity to beat the Axis powers. Recessions have the opposite effect: they leave working people with deteriorating job options, making the prospect of being replaced by a robot seem much worse.

Unlike the war, Covid-19 requires people to stay at home. Furthermore, with jobless claims rising at record rates and the world economy shrinking, automation anxiety looks set to witness a revival — and with good reason. Coronavirus is likely to accelerate automation. 

For one, companies strive to cut costs during downturns. An influential study by Nir Jaimovich and Henry Siu showed that routine jobs that can easily be automated vanished with the 2008 financial crisis and didn’t come back, contributing to a jobless recovery.

Consumers also tend to “trade down” during a recession, opting for cheaper and less labour-intensive goods and services. For example, if US consumers switch from Whole Foods (which employs six workers per $1m of sales) to Sam’s Club (which employs two workers per $1m of sales), they increase the level of automation in the economy.

Even after Covid-19 passes, companies will want to pandemic-proof their operations. Ecommerce has been boosted by social distancing, but pressure has meanwhile mounted on online retailers as their warehouses have become overcrowded. Indeed, workers at Amazon warehouses in the US and Europe have protested. Warehouses need such workers as picking stock remains a highly manual process. But robotic hands are becoming more dexterous by the day, and algorithms are now capable of better distinguishing between objects, enabling automation.

Coronavirus may change consumer behaviour, too. During the 1918 pandemic, fear of contracting Spanish flu radically altered social interactions via a long-lasting reduction in trust. Today, consumers may similarly prefer automated services to face-to-face interactions for some time to come.

Finally, while the pandemic is reigniting efforts by the Trump administration to reduce its dependence on China and to reshore the production of drugs and medical goods, such a move is unlikely to boost job creation. Besides the enormous disruptions it would cause to global supply chains, it would only accelerate automation as companies seek to make up for higher labour costs. Even in China, about 12.5m production jobs vanished from 2013 to 2017 as robot use accelerated. If US-China tensions escalate further, the decoupling Donald Trump has advocated could happen on a giant scale, with a growing share of manufacturing done in automated factories in the west.

How will working people fare in such a world? Most jobs likely to be automated are of lower pay. In 2016, based on our research, then-president Barack Obama’s Council of Economic Advisers estimated that 83 per cent of workers in occupations that paid less than $20 an hour were at high risk of being replaced, while the corresponding figure for workers in occupations that paid more than $40 an hour was only 4 per cent.

High-income jobs are not only harder to automate, they are also more pandemic-proof. Of 483 occupations analysed in a forthcoming report by Citi and the Oxford Martin School, 113 of them — accounting for 52 per cent of the US workforce — could be performed remotely. By contrast, high-income earners are five times more likely to be able to work during the pandemic. While the virus doesn’t discriminate, automation and digital technologies are exacerbating social cleavages and could be a source of unrest for years to come.

In 19th-century England, Luddites smashed machines, fearing for their jobs. Their riots were ended by force — the army used against them was larger than that against Napoleon in 1808.

Automation helps drive long-term prosperity. But in a world where workers are also voters, Luddite efforts to avoid its short-term costs can also deny the benefits, an effect I call the “technology trap”. Today, governments have made huge efforts to battle the economic consequences of the virus. But when job subsidies and other measures expire, automation still looks set to accelerate. In a time of cash-strapped consumers and growing robotics, some form of basic income may be necessary.