Study: Automation Drives Income Inequality | MIT News

When you use self checkouts in supermarkets and pharmacies, you’re probably not, with all due respect, doing a better job of packing up your purchases than the cashiers once did. Automation makes bagging less expensive for large retail chains.

“If you introduce self-checkout kiosks, productivity won’t change much,” says MIT economist Daron Acemoglu. However, in terms of lost wages for employees, he adds: “It will have quite broad distributional effects, especially for low-skilled service workers. It’s a work-shifting device, rather than a productivity-boosting device.

A recently released study co-authored by Acemoglu quantifies the extent to which automation has contributed to income inequality in the US simply by replacing workers with technology, whether it’s self-checkout, call center systems, chain tech mount or other devices . Over the past four decades, the income gap between educated and educated workers has grown significantly; the study finds that automation accounts for more than half of that increase.

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

The paper, “Tasks, Automation, and the Rise in US Wage Inequality,” was published in Econometric. The authors are Acemoglu, an institute professor at MIT, and Pascual Restrepo PhD ’16, assistant professor of economics at Boston University.

Lots of “so-so automation”

Since 1980 in the United States, the inflation-adjusted incomes of those with college and postgraduate degrees have increased substantially, while the inflation-adjusted earnings of men without high school degrees have decreased by 15%.

How much of this change is due to automation? Growing income inequality could also result from, among other things, the declining prevalence of unions, market concentration resulting in a lack of competition for labor, or other types of technological change.

To conduct the study, Acemoglu and Restrepo used statistics from the US Bureau of Economic Analysis on the extent to which human labor was used in 49 industries from 1987 to 2016, as well as data on machinery and software adopted during that period. The scholars also used data they had previously collected on robot adoption in the United States from 1993 to 2014. In previous studies, Acemoglu and Restrepo found that robots alone replaced a substantial number of workers in the United States, helped some companies dominate their industries and contributed to inequality.

At the same time, the researchers used US Census Bureau metrics, including data from the American Community Survey, to track worker outcomes during this period for approximately 500 demographic subgroups, broken down by gender, education, age, race, and ethnicity and immigration status, examining employment, inflation-adjusted hourly wages and more, from 1980 to 2016. By examining the links between changes in business practices and changes in labor market outcomes, the study can estimate the impact that automation has had on workers.

Ultimately, Acemoglu and Restrepo conclude that the effects have been profound. Since 1980, for example, they estimate that automation has reduced the wages of men without a high school diploma by 8.8% and of women without a high school diploma by 2.3%, adjusted for inflation.

A central conceptual point, says Acemoglu, is that automation should be viewed differently from other forms of innovation, with its distinct effects on the workplace, and not just as part of a larger trend towards implementing technology in everyday life in general.

Consider those self-checkout kiosks again. Acemoglu calls these types of tools “so-so technology” or “so-so automation”, because of the trade-offs they contain: such innovations are good for corporate bottom line, bad for employees in the service sector, and not hugely important in terms of overall productivity gains, the true indicator of an innovation that can improve our overall quality of life.

“Technological change that creates or increases the productivity of the industry, or the productivity of a type of job, creates [those] large productivity gains but does not have huge distributional effects,” says Acemoglu. “Conversely, automation creates very large distributional effects and may not have large effects on productivity.”

A new perspective on the big picture

The results occupy a distinctive place in the literature on automation and work. Some popular accounts of the technology have predicted near-total job losses in the future. Alternatively, many scholars have developed a more nuanced picture, in which technology disproportionately benefits highly educated workers, but also produces significant complementarities between high-tech tools and manpower.

The current study differs in at least an extent from the latter picture, presenting a starker perspective in which automation reduces earning power for workers and potentially reduces the extent to which policy solutions – more bargaining power for workers, less market concentration – could mitigate the harmful effects of automation on wages.

“These are controversial findings in the sense that they imply a much larger effect for automation than anyone else has thought, and they also imply less explanatory power for other [factors]”, says Acemoglu.

However, he adds, in an effort to identify the drivers of income inequality, the study “doesn’t completely rule out other non-tech theories. Furthermore, the pace of automation is often influenced by various institutional factors, including the bargaining power of labour.”

Labor economists say the study is an important addition to the literature on automation, labor and inequality and should be considered in future discussions of these issues.

“Acemoglu and Restrepo’s paper proposes an elegant new theoretical framework for understanding the potentially complex effects of technical change on the aggregate structure of wages,” says Patrick Kline, professor of economics at the University of California, Berkeley. “Their empirical finding that automation has been the dominant factor driving wage dispersion in the United States since 1980 is intriguing and looks set to reignite the debate about the relative roles of technical change and labor market institutions in generating inequality salary”.

For their part, in the paper Acemoglu and Restrepo identify multiple directions for future research. This includes studying the response of both business and labor to increased automation over time; the quantitative effects of job-creating technologies; and industry competition between companies that have rapidly adopted automation and those that haven’t.

The research was supported in part by Google, the Hewlett Foundation, Microsoft, the National Science Foundation, Schmidt Sciences, the Sloan Foundation and the Smith Richardson Foundation.


Leave a Reply

%d bloggers like this: