Automation, productivity, and wages

In an article called Delayed expectations: Automation, productivity and wages for Barclays’ Our Insights, Managing Director and Head of Macro Research Ajay Rajadhyaksha and Director and Head of FX and EM Macro Strategy, Americas, Aroop Chatterjee outline just how much our (society as a whole, economists, journalists) expectations for the impact of automation differ from reality.

While people fear that robots will take over human jobs, with a devastating impact for individuals and society, unemployment, wage growth, and productivity data paint a different picture. One that seems to be a recurring pattern throughout human history: as technology advances exponentially, we continue to overestimate its rate of adoption and overall impact on the economy.

We have a long way to go until hard automation, technology that automates 100% of an individual’s job and replaces the human worker, becomes widely available across industries. In the first decade after introduction, soft automation, where only parts of a job are automated, is achievable. In other words, robots are not capable to replace humans yet. And history shows that the introduction of cars killed horse-related jobs, but also created a plethora of new jobs in car production, operation, and maintenance.

Soft automation will continue to result in reduced wages as it eases the learning curves of professions and automates specific tasks within professions. If a profession requires less skill and less specialization, wages go down. This is the reality for truck drivers, industry workers, and even accountants. Jobs will become less specialized, competition on the market will increase (as more individuals are able to perform the functions), and wages will go down.

Productivity is not likely to dramatically increase, despite the introduction of disruptive tech like the Industrial Internet of Things (IIoT) and Machine Learning (ML). Most of the great technological leaps happened in the Second and Third Industrial Revolutions. Today, we’re running on legacy infrastructure that’s permitting us from achieving major leaps in productivity.