This piece was originally posted in SAIS Perspectives.
The Future of Work is all the rage—the latest World Bank World Development Report is dedicated to the issue, the World Economic Forum and McKinsey have issued multiple reports, and "your job is being stolen by a robot" articles are being churned out at a rate that suggests IBM’s Watson has been reprogrammed to deliver 50 think pieces a day. We’ve been here before, of course: dire warnings of a machine age of mass unemployment have been around for at least a century. Is this time different? I don’t think there’s compelling evidence yet to suggest it is. That said, technology change has long interacted with policy choices to influence who gains from production both nationally and globally, so even if this is business as usual, policy makers should respond.
There are lots of reasons to think we aren’t in the midst of a revolutionary worldwide shift in employment patterns. Looking at the US, job churn (i.e., a measure of people switching employment) is low, productivity is inching up, and employment overall is climbing. Developing countries haven’t suddenly stopped adding jobs in industry (though, admittedly, they haven’t expanded as fast as you might have expected over the past decades) and services employment keeps on growing. We’re still in a period of global income convergence that has been running for two decades. If robots and AI were radically altering work to the disadvantage of low and mid-skilled workers, we’d expect none of those things to be true. And it is worth remembering that the key reason that poor countries are poorer than rich countries is because they see less extensive use of productivity-enhancing technologies (e.g., hand threshing rather than combine harvesters, kerosene lamps rather than LEDs). We should all want more automation and changing employment patterns in Tanzania and Togo, not less.
Another thing that hasn’t been changing as rapidly as you might have expected during a robot revolution is political views. There’s a standard response to the future of jobs question—one that I use, too—that automation and AI make whatever policy choices we thought were right before even more right. Among these choices exist options such as redistribution though social programs, prevention of monopolies, and regulation of things like labor standards.
To be fair, some seers in Silicon Valley have suggested something comparatively new: that the coming jobs apocalypse calls for the introduction of a universal basic income (UBI)—everyone gets a monthly check from the government enough to support a minimum standard of living. The idea has the virtue of simplicity and the vice of immense cost. $15 a day would leave people a little below the poverty line; provide that to every American and it comes to around $1.8 trillion a year (compare that to about $0.18 trillion in means-tested cash assistance programs currently provided by the US government). Roll out a less generous $5 global basic income to deal with the potential global consequences and that would come to $12.8 trillion, give or take—85 times current global aid budgets. It would be wonderful, as well as incredible, if the world were so egalitarian to fund such an exercise; since in practice it is not, we should aim to strengthen our existing policy tools instead.
That said, perhaps automation and AI will pick up speed over the next few years—the kinks in driverless cars will be worked out, factories will require fewer and fewer semi-skilled assembly workers and tasks from writing legal agreements to reading MRIs will increasingly be done by computers. And robots or no, existing trends in national inequality and the still-yawning gap in global incomes do demand policy responses. People are worried that good manufacturing jobs are being replaced by bad services jobs, but there is nothing that preordains services jobs to be bad or manufacturing jobs to be good (see: The Jungle, Rana Plaza).
Decent work has required, and will always require, policy action—not least labor laws and guaranteeing worker rights—whatever sectors of the economy it is in. Inequality started expanding in the US when labor laws were weakened, top tax rates were cut, banking was deregulated and competition policy was largely abandoned. Reversing those policy choices would be a great way to claw back inequity. At the global level, easing trade in services, not least through reduced constraints on the movement of people, could provide a powerful additional path to growth beyond manufactured exports. Robots are no reason to throw up hands in despair; productivity gains—if they emerge—are a reason to celebrate, and the task of distributing those gains fairly demands the same old policy responses we’ve delivered so semi-competently all these years.