Automation is driving huge increases in wage inequality

The wage gap between the most and least educated US citizens has widened sharply over the past 40 years, and new MIT research finds that more than half of this disparity can be attributed to a single factor: automation. This is a bad omen.

The value of the gross domestic product of the United States has increased from US$6.82 trillion in 1980 to over $20 trillion in 2022. But with almost three times the cake for everyone, not everyone ended up with more on their plate.

According to the MIT economist David Authorthings have been radically different in the last 40 years than in the period between 1963 and 1972, when “wages rose robustly and evenly across all education gender groups.”

While real wages have risen sharply for graduates since 1980, American men without high school degrees earned 15% less in 2016 than they did in 1980, adjusted for inflation. The story is similar in the UK and Germany.

Americans' wages rose together until around 1980, when the gap in wage growth between the more and less educated began to split completely.
Americans’ wages rose together until around 1980, when the gap in wage growth between the more and less educated began to split completely.


A new study by two other MIT economists, Daron Acemoglu and Pascual Restrepo, questions this relatively recent shift toward greater inequality using a new framework that analyzes tasks and assigns them to different types of labor and capital. Simply put, this new model makes it relatively easy to link pay changes in a particular group to task shifting, that is, when machines start doing someone’s job.

It has long been obvious that simple, repetitive jobs do not have a bright future, and in fact, it is this type of unskilled work that has been most easily replaced by automation. But the MIT team argues that, even when controlling for other factors such as declining union participation, the impact of automation on the wages of affected groups is considerably worse than previously thought.

“This single variable… explains 50 to 70 percent of the changes or variation between group inequality from 1980 to around 2016,” Acemoglu said in a news release. “These are controversial findings in the sense that they imply a much larger effect for automation than anyone has thought, and they also imply less explanatory power for other [factors].”

The study also draws a distinction between automation that actually increases overall productivity—for example, moving to a robotic factory production line that makes things faster and with fewer errors than human workers—and automation that eliminates jobs. and saves money for business owners without any increase. in productivity: the perfect example is the loss of supermarket cashier jobs to self-checkout machines, which don’t do the job much better or faster.

This kind of “regular automation,” as the authors describe it, is an example of the kind of automation that does little more than funnel money down the social strata, away from unskilled workers. It favors one group over another, with no broader benefits to society to offset this increase in inequality.

Why should the better educated among us care about this, beyond simple empathy for our fellow man? Well, the last 40 years it’s been just a tasting spoon. The double ice cream of automation is being served as we speak, in a technological convergence of artificial intelligence, big data, robotics, connectivity, energy storage, and a host of other factors that promise to simplify the job of automating a much broader range of jobs.

Capital often sees workers as black-box algorithms: given input A, it expects output B. But digital algorithms are getting smarter very quickly, while human brains have arguably done little more than atrophy. in the last hundreds of thousands of years. Today, the less educated find their purchasing power crippled by automation and decimated by inflation. Tomorrow, it will be high school graduates and then college graduates, at an accelerated rate.

“The pace of automation is often influenced by various institutional factors,” Acemoglu said, “including the bargaining power of workers.” But therein lies the problem; What is the bargaining power of labor when capital does not want it? Humanity is heading into uncharted waters in this regard. Throughout history, the rich and the poor have reached a difficult agreement that took everyone on a boat trip, because the rich have always needed the poor to do the work. Labor has held capital accountable for its worst excesses by occasionally pulling its one great lever: band together and stop labor. Some have even argued that it is the value of human labor, rather than something more esoteric and maudlin, that underpins the modern Western notion of human rights.

And so we move towards a future where more and more kinds of human labor become less and less valuable. Without some unprecedented structural changes, it certainly looks like the effects will be brutal.

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

The study is open access in the journal econometric.

Font: MIT

Leave a Reply

Your email address will not be published.