What’s the Best Way To Help Low-Income Workers? Automate Low-Income Jobs.

As we celebrate Labor Day, reducing unemployment and getting the COVID-impacted economy back to some semblance of normality is clearly the top economic task. But when that is done the economy will still face a critical labor market problem: too many workers earning too little. A recent Brookings study found that 44 percent of American adults workers make very little, with median annual earnings of just $18,000.

There are two broad approaches to addressing this challenge. The first, and most popular among advocates and policy makers, is to transfer more societal resources to these workers. They would still be working in low-wage jobs, they would just be getting more money. A leading proposal is to increase the earned income tax credit. A related concept is wage subsidies, where the employer gets public funding in exchange for increasing wages. There are three key problems with this approach. The first is that it reduces spending available for key areas of national investment, like boosting R&D. The second is that it is not likely to be big enough to make a significant difference: the EITC provides only around $3,200 for a family with children. The third is that it makes it easier for companies to stay on the “low-road” path of low wages, low training and low investment.

An alternative is to boost the minimum wage and/or increase benefits, including by requiring health insurance and other benefits. This is superior to the employer subsidy approach for three reasons. First, it doesn’t impact the federal budget.  Second, depending on the level of intervention, it would provide a larger benefit than the EITC.  And most importantly, making employers pay more, whether directly or indirectly, it provides an incentive for employers to boost productivity. Indeed, the evidence is clear that low wages reduce capital investment and other measures employers can take to boost productivity.

Some propose an alternative: investing in more training, in part to boost mobility. In fact this is seen as a panacea by both sides of the aisle. The Aspen Institute, for example, proposes a “new capitalism” which would have employers invest in the “long-term development of their workers,” rather than short-term profits. Few are against more “human capital formation.” But mobility to what? It’s not as if there are millions of high-wage jobs waiting for people working at fast food restaurants to get their college degree. Unless we create more good jobs, all that mobility strategies will do is let some people move to higher paid jobs while other people will move to their lower-paid jobs.

This gets to the second strategy, creating more higher paying jobs. If the economy can’t create more higher-wage jobs, all the programs and policies in the world to improve low-wage jobs will be a band-aid at best. But to the extent policy advocates and policy makers focus on good jobs, their focus is on increasing the absolute number of good jobs, rather than increasing the ratio of good-to-bad. Nationally government can do little to increase the number of good jobs, beyond reducing the trade deficit (traded sector  jobs pay about 10 percent more than non-traded jobs). Beyond that there isn’t much to be done, especially because the addition of more good jobs depends on increased purchasing power to support those jobs.

This means that any effective strategy needs to be focused on eliminating more low-wage jobs, usually through automation. If an employer automates a low wage job, the price of the good or service the worker used to produce falls, expanding economy-wide purchasing power. That added money gets spent, creating demand for jobs at all income levels, from doctors, to broadband installers, to teachers. In other words, if we could dramatically boost the productivity of low-wage jobs the result would be fewer low-wage jobs and more middle and higher-wage jobs.

Fortunately, the next wave of technological innovation, grounded in artificial intelligence, internet of things, and robotics, is likely to have a higher impact on lower wage jobs. As ITIF has estimated, the correlation between the average wage of an occupation and its risk of automation over the next decade is negative and quite large (around -0.55).

Amazingly, many actually advocate doing the opposite as way to help low wage workers. The MIT Task Force on the Future of Work proposes reducing incentives for employers to invest in automation and for government to only support automation technology that doesn’t replace workers (what they call “complementary technologies”). This channels the currently faddish view that automation is bad for workers and that government should actually slow automation. Nothing could be farther than the truth.

In fact, a strategy consisting of reducing the incentives for employers to take the low road (including limiting low-wage immigration and boosting the minimum wage), coupled with policies to support  automation, including tax incentives for new capital equipment, and stronger policies and programs to help workers make occupational transitions could make a real difference in shifting the U.S. economy to one with a much larger share of good, middle-class jobs.