October 20, 2019
Automation Risk, Trade Risk, and Public Policy
The debate on trade and automation was on prime display during the democratic presidential debate this week. There is a lot packed into this discussion, from trade policy and taxation of capital to the role of place-based economic development efforts and the design of pre-K, high school and college curriculum. My colleagues and I at Ball State have written widely on these issues, and I am familiar with the technical details that underlie this debate. It isn’t going away, so it is good to focus on facts and risks, not hunches and soundbites.
Over the past 50 years, the US has been a job-creation machine, adding some 82 million jobs. Of this job growth, 88 million, or about 107 percent, have been jobs outside of factories. Yes, I did the math right. We have 6 million fewer factory jobs since 1969.
The biggest loss of jobs happened from about 2007 to the end of the Great Recession. That was a time of rapid new imports from China, and it was also a time when factories were experiencing significant productivity growth. Both of these factories shifted the need for workers in U.S. factories, leading to large job losses.
Productivity growth, rather than imports, were the direct cause of most of these job losses. However, a significant share of the productivity growth can surely be attributed to factories responding to the threat of losing market share to China and other foreign competitors. In this case, the threat of imports accelerated productivity gains that would have come anyway, but at a more gradual and measured pace.
The change in productivity had many sources. Some of it was automation and robotics directly. One recent paper that Senator Warren referred to claims that automation alone was responsible for about half the job losses that were claimed by trade. But, other factors made our factories more productive. Most importantly, workers got better. By 2007, half of all factory workers in the US had been to college or had an associate’s degree. Given the modest hiring in factories over the last decade, that means nearly all the net job growth in factories nationwide went to those with more than a high school diploma.
Factories adopted technology in the 1990s and 2000s, which allowed them to re-organize production. Computer technologies reduced warehouse space, and digitized inventories made factories leaner. The adoption of statistical process controls caught costly production mistakes early, saving time, money and people.
Robots and robot-like equipment also boosted productivity, even if the burst of new purchases don’t appear in the manufacturing investment data. The reason for this is that a robot is really a regular old machine with a new brain. The new brain involves sensors and a computer that are, in reality, quite cheap. What makes them expensive is the software, or artificial intelligence, that operates them. It costs little to replace the brawn of a Homo Sapiens, but even the simplest learning of tasks is very costly. Artificial Intelligence is just software that permits a machine to learn a new task, using some forms of sensors or perhaps a person guiding them through a process with a game controller.
Artificial intelligence and robotics will scarcely appear in the investment data of a company. It’s usually just outsourcing of some new service, like software maintenance. One study even found a large number of factory job losses over the past decade were due to companies changing their industrial codes from manufacturing to services. I think this is an example of that phenomenon.
The debate about the share of jobs lost to trade or technology won’t go away, but mostly that is because it offers convenient political rhetoric and the falsehood of a quick remedy. The current trade war is a failed policy fiasco, as would be an extension of it by a democratic president. I know it’s unpopular to say, but the badness of a policy is indifferent to the political affiliation of its supporters.
Still, we should focus on automation, even if we had a way to reverse trade with China. That is simply that the China Shock (as it has been termed) is a one-time event. There are no more billion person economies waiting around to trade with us. The future risk is that of automation and the disruptions it causes.
Increased productivity growth and automation will create more jobs than it destroys, but the new jobs will be in different places, and require different skills than the old jobs. We could do some things to help keep more jobs in struggling places, and perhaps we should. But, the biggest risk of automation accrues to the least educated workers. Their jobs are inherently easier to automate, and they are less likely to possess the skills to retrain, or do other work.
Higher automation risk requires workers have better educational foundation. That means more investment in early education, pre-k through perhaps 3rd grade. It means stronger math and literacy skills in middle school and more academic focus in high schools. It also means more students should be attending, and then more graduating from colleges.
Education in fundamentals is about the only way to insulate individual workers from automation risk. But, against this backdrop, Indiana has eased graduation requirements, and cut the share of GDP spent on both K-12 and higher education. We have seen a major drop in the number and share of students pursuing a 4-year degree, and all our major universities now have fewer in-state students than we did a decade ago, despite a growth in high school graduates. Indiana is the single most at-risk state for automation-related job losses, and we are busy making it worse.
About the Author
Educational Attainment, the 21st Century Fund and the Future of SchoolingIndiana ranks 42nd in educational attainment.
Big Savings for Ending Prevailing WageMy statistical models show that repealing state prevailing wage laws save taxpayers money.
Re-Thinking Economic Development A large share of the most mobile families—perhaps half—no longer need to live near where they work.
Money Illusion and InflationPrice fluctuation could cause inflation to last longer, but it didn’t cause the inflation, it simply extends the pain.View archives