Automation risks costing people their health as well as their jobs

Getty Images

Unease over the role of automation-related job losses has risen in recent years, spawning academic conferences, research studies and a host of news stories. To be sure, productivity growth through better automation and work processes should be a welcomed feature of any economy.

There is a justified concern, however, for workers who find themselves unable to work due to obsolete skills. That worry is the basis for significant investments in workforce training and skill enhancements in every state, but what if there’s more to the problem? Does the risk of automation-related job losses effect more than labor market outcomes?

{mosads}That question is central to a new study on automation-related employment risk and health outcomes at the county level. In this work, colleagues at Villanova and Ball State Universities investigated whether the higher prevalence of workers exposed to automation-risk resulted in different health-care outcomes.


To do this, we used County Health Rankings, matched a study on automation risk from Oxford University to the county level and controlled for other conditions which might bias our results, including:

  • race,
  • age,
  • existing income inequality,
  • educational attainment,
  • access to health care and exercise opportunities,
  • the effect of job losses due to offshoring risk from one of our previous studies,
  • household income and
  • other economic measures all at the county level.

We also included growth in imports from China since 1990 and the average risk of automation of adjacent counties.

This is the kind of very technical study that professors regularly perform, but what was surprising to all of us is the large and statistically meaningful role automation-related job-loss risk plays on health-care outcomes across the board.

Our study found that a 10-percent increase in automated job-loss risk in a county increased the average county ranking of poor health by as much as 1 percent. Likewise, an equally large increase in the risk of automation-related job losses increased self-reported levels of frequent physical and mental health distress.

The financial cost of this distress due to absenteeism is large:

  • a 10-percent increase in risk of automation-related job losses would increase costs by $24 million to $174 million due to increase in prevalence of poor or fair health;
  • $6 million to $40 million due to increase in physical distress; and
  • $7 to $47 million due to increase in mental distress.

In a typical county, that means the impact of automation-related job-loss risk would range from $37 million to over $250 million in losses due to poorer health outcomes alone. Beyond the obvious, this is important because between about a quarter and a half of all jobs at risk at the county level are being lost to automation by technologies already in commercial use.

This research and an abundance of studies that convincingly link economic worries to lower levels of health make us confident that there is a causal link between automation-related employment risk and poorer average health in a county.

Data limitations prevent us from answering questions about individual health-care outcomes, however. That will have to wait for a later study.

There is also some useful policy inference from this study. The first is that automation-related productivity growth is still a net benefit. Whatever we might do to limit the costs of automation should not interfere with productivity growth across the economy.

The reality is that we need workers to fill jobs today that are at risk of automation in the future. As such, policy should focus ensuring that those workers have the technical, transferable skills needed to retrain after automation and the coping skills necessary to manage the stress of job loss.

Indeed, the most beneficial policy intervention would be better fundamental education at the K-12 level, where the skills that allow workers to retrain are first taught, and more comprehensive skill training for displaced workers.

Finally, it is important to restate that while productivity change is good for the economy and households generally, there is reason to imagine that the turbulence to some households is large, persistent and extends beyond labor market outcomes.

This study adds to our understanding of that issue, while acknowledging that productivity growth lies at the heart of a growing economy.

Michael Hicks, Emily Wornell and Srikant Devaraj are all professors at the Center for Business and Economic Research at Ball State University. Pankaj Patel is the Frank J. and Jane E. Ryan endowed chair in strategy and innovation at Villanova University.

Tags Automation Economic globalization Economic inequality economy Risk Social inequality Structure Technological unemployment Unemployment

More Finance News

See All
See all Hill.TV See all Video

Most Popular

Load more


See all Video