Backgrounder: Policy Responses to Technological Unemployment
More and more economists are forecasting that automation will displace massive numbers of workers from their jobs. In response, there is now extensive literature and public debate as to how the United States ought to react to automation and manage its effects, if it should at all. This post surveys the various ideas governments, unions, and individuals have proposed to address the labor displacement that increasing automation may cause, and the extent to which those ideas have been implemented in the past and are being implemented in the present.
There are two broad categories of proposals surveyed: first, proposals to prevent labor displacement by preventing the diffusion of automation technology; second, proposals that accept any labor displacement that may occur because of automation, and instead deal with the effects of that labor displacement.
Preventing Labor Displacement by Preventing Automation
Outright bans are the bluntest method of slowing, or stopping altogether, the proliferation of labor-saving technology. Earlier this year, India’s top transportation official stated that the government would ban self-driving cars from operating in India to protect people’s jobs. India’s stance stands in contrast to that of U.S. regulators, who have generally expressed support for the development of driverless technology, though nearly 3% of the American workforce make their living driving taxis, buses, vans, and trucks.
This enthusiasm, however, is not shared by all Americans. San Francisco City Supervisor Norman Yee has proposed banning sidewalk delivery robots, primarily for safety reasons, but also out of concern that the robots will take away delivery jobs. The Upstate Transportation Association, a lobbying organization for professional drivers in upstate New York, has urged the state of New York to ban self-driving cars from the state’s roads for 50 years.
Market-based mechanisms, which work by increasing the opportunity cost of automation, are more commonly proposed than bans. So-called “robot taxes” are proposals to either directly tax automated machinery or, more commonly, to tax the capital gains accrued by corporations from the use of automated machinery that has a labor-displacing effect. In a bill that was later rejected, Member of European Parliament Mady Delvaux proposed taxing corporations for reliance on automated equipment as a means of protecting workers. Bill Gates has supported proposals like Delvaux’s. Gates’s endorsement of robot taxes recently led San Francisco Supervisor Jane Kim to create the nascent Jobs for the Future Fund, which proposes that “as workers are displaced,  companies continue to pay a portion of the lost tax into a fund that can then be used for education, retraining and targeted investments in new industries.”
Instead of increasing the costs of automation, some have proposed decreasing the costs of labor to make automating less attractive of an option. Automation has been cited as a reason to push back against movements for greater worker protections, including the Fight for $15. Andy Puzder stated that if he had been successfully appointed Secretary of Labor, he would have prioritized making entry level jobs accessible by keeping wages low.
The argument for limiting worker entitlements so as to save their jobs is not limited to conservatives, either. New York University law professor Cynthia Estlund has proposed that policymakers shift the economic burden of certain worker entitlements away from the employer to other actors, like the state. She gives health insurance as an example — employers do not, as a practical matter, have to be the entity that provides health insurance to employees. And if they were not, the additional “tax” born by employers from hiring employees would be reduced, encouraging employers to use employees in lieu of independent contractors or automated machinery.
Dealing with Automation’s Effects
In addition to policies that would prevent or slow the diffusion of automation, there are policies that prepare workers for its eventual effects.
The most modest of these is retraining workers into less automatable positions. In order to be a viable new job for the worker, the job that the worker is retraining for must be both in high demand and not vulnerable to automation soon. Otherwise, the worker will later have to expend considerable investment to retrain, yet again, for a new job. Tech jobs are often recommended to workers displaced from automation, since demand is high and entry barriers are low.
With so few workers today in a union, a mass retraining program would have to be publicly delivered in order to reach most of the workers whose jobs are at risk from automation. One model for how a public retraining program would work is the Trade Adjustment Assistance program, which allows people who can demonstrate that they lost their jobs as a consequence of American trade policy to receive free schooling and cash for up to two and a half years, or an apprenticeship paid for by the government.
Among people who believe that automation’s impact on labor markets will be quick and far-reaching, however, retraining is an insufficient response. These individuals instead focus on ensuring that displaced workers have an acceptable standard of living, regardless of their employment status. Labor economist Richard Freeman argues that “workers need to own part of the capital stock that substitutes for them.” To this end, economist Noah Smith has suggested that workers receive a citizen’s stock portfolio in companies that rely heavily on automation.
But the most popular proposal to address the effects of automation is, by far, universal basic income. Though significant variation in the idea exists, at its core, UBI is the promise of a regular, unconditional sum of money distributed from the government to people for them to spend on whatever they like. Recently, OnLabor featured posts from former SEIU president Andy Stern and from law professor Brishen Rogers discussing the merits of UBI from a labor-centered perspective. There is extensive literature on UBI, and the idea has received favorable attention from both the left and the right. Perhaps the biggest political barrier for UBI right now is cost. Cost estimates range from the hundreds of billions to the trillions, depending mostly on the extent to which UBI supplants existing welfare expenditures.
Prior attempts and experiments with UBI have been mostly concerned with its feasibility as an alternative to the welfare state. The threat of automation plays a bigger role in the present interest in UBI. In his book Raising the Floor, Andy Stern says he came to support universal basic income partly out of concern with automation’s impact on workers. And Y Combinator, the largest and most successful startup accelerator in Silicon Valley, is running a randomized controlled trial testing the effect of unconditionally giving $1000 per month for five years to U.S. citizens to “advanc[e] the debate about social spending and the future of work.”
It remains unresolved whether automation will actually cause the mass displacement of workers. In any case, the uncertainty of the future has encouraged, not deterred, people from thinking of ways to manage automation and its implications. In the event that the risks of automation are real, measures that seek to deter automation and measures that seek to deal directly with its consequences will both likely be necessary.
This is a summary of a paper which can be accessed here.