Over the past 40 years, income inequality has been rising rapidly in several countries — especially in the US. While there are many factors contributing to this, one single variable, automation, accounts for over half of this increase.

If you’ve ever used a self-checkout service at a supermarket, you’ve probably noticed that you’re not really faster than a clerk. You’re probably not saving any time, and you’re not really any better off than with a human clerk. The only party benefiting from this is the store, which saves money as they don’t need to pay another salary.

In economic terms, you’re not increasing the overall productivity or capital. You’re not really adding any resources to the systems.

“If you introduce self-checkout kiosks, it’s not going to change productivity all that much,” says MIT economist Daron Acemoglu. However, in terms of lost wages for employees, he adds, “It’s going to have fairly large distributional effects, especially for low-skill service workers. It’s a labor-shifting device, rather than a productivity-increasing device.”

This is just one example of how automation is affecting jobs. When we think about automation, we usually conjure up images of fancy robots and complex machinery, but oftentimes, something as simple as a self-checkout-service can make a big difference, especially when it comes to so-called “low-skill jobs”. So Acemoglu wanted to see what kind of effect this is having on inequality.

In the US, inequality has been rising steadily in the past decades. While the salaries of those with low education (without high school degrees) have actually decreased, the salaries of those with college degrees have been substantially rising. Furthermore, the rich are getting richer and the inequality gap has been steadily increasing.

This inequality is owed to many things. Policy and labor protection is one aspect, and the decline of labor unions is another. The market could itself be shifting — there’s a million and one factors that could be driving this shift. But remarkably, Acemoglu found that by far the biggest impact comes from automation.

“This single one variable … explains 50 to 70 percent of the changes or variation between group inequality from 1980 to about 2016,” the economist says.

Acemoglu and Pascual Restrepo, an assistant professor of economics at Boston University, gathered information on how much human labor and how much machinery and software was used between 1987 and 2016. They mixed that with data previously compiled about the adoption of robots gathered from 1993 to 2014, as well as US Census Bureau metrics. These metrics gather data for 500 demographic subgroups, sorted by gender, education, age, race and ethnicity, and immigration status, while incorporating data on employment, inflation-adjusted hourly wages, and more, from 1980 to 2016.

The researchers then looked at how the labor market changed as automation was introduced and drew conclusions about how one influenced the other.

The findings show that automation has had quite a big effect. For instance, since 1980, the wages of men without a high school diploma have dropped by 8.8%, and wages of women without a high school diploma have dropped by 2.3% (adjusted for inflation).

So-so automation

From the perspective of business owners, automation is great. But from the perspective of the economy as a whole, automation doesn’t really do all that much. Unlike other forms of innovation, it doesn’t really bring any new advantages to the table, it just makes companies more money at the expense of workers in “low-skill jobs.” Acemoglu calls this “so-so automation.” 

“Technological change that creates or increases industry productivity, or productivity of one type of labor, creates [those] large productivity gains but does not have huge distributional effects,” Acemoglu says. “In contrast, automation creates very large distributional effects and may not have big productivity effects.”

However, the two economists don’t advocate giving up on technology. Rather, they say, it’s important to consider and account for the effects of automation on the labor market, especially on a policy level, and prepare for the effects this will have in terms of inequality.

It’s not the first time Acemoglu has looked at how automation is impacting the market. For instance, one previous study found that robots by themselves are replacing a substantial number of workers and actively contributing to inequality. Other researchers have also found that robots are most likely to take over jobs that pay $20 an hour or less and since we have industry-specific estimates on what jobs are most likely to be automated, it makes sense to prepare social and labor policies that would protect the workers most at risk and ensure that income inequality doesn’t surge even more.

Income inequality is a problem for everyone in our entire society, not just for the ones directly affected. The effects of income inequality range from lower population-wide satisfaction and happiness to a lower level of economic growth and more health and social problems.

The study was published in the journal Econometrica.