Whether you look at the first Industrial Revolution or the fourth – the 21st century’s own technology-led transformation – automation has always been regarded with trepidation.

The assumption is that machines take away jobs from people, causing unemployment and economic inequality. To workers in the 19th century, the threat came from the spinning jenny and the steam engine. Today, people worry about artificial intelligence and robots rendering workers obsolete.

There is even a body of scientific research which suggests such technology pessimism may be justified; that computers will put at risk a large share of occupations, especially mid-wage roles; and that they will consequently polarize employment between low-end and very highly paid jobs.

James Bessen, an economist at Boston University has been testing these claims by exploring the relationships between technology and occupations. He comes to two surprising conclusions.

The first is that occupations that use computers grow faster, not slower – and there is no sign of technology causing large scale unemployment or polarization.

The second insight is that instead of job losses, there is a resulting need for new skills: “New technology can increase demand for an occupation, offsetting putative job losses.”

 Figure 4
Image: James Bessen, Boston University School of Law

Moving up the value chain

In his paper, “How Computer Automation Affects Occupations”, Bessen points out that the manufacturing workforce grew from less than 12% in 1820 to 26% by 1920 despite automation. Similarly, numbers of cashiers have increased since the introduction of barcode scanners in the 1980s. While Internet sales now account for over 7% of retail turnover, the number of people in sales jobs has grown since 2000.

His own research finds that overall, the net impact of computers on the number of jobs available is negligible.

Employment grows faster in jobs where computers are used more but this trend is offset by the fact that computer-based roles replace other occupations. Computers boost the number of highly paid job roles requiring more educated workers - to the detriment of low-paid jobs. In a nutshell, this means that jobs don’t disappear, they simply move up the skills and wage ladder.

For workers to move up the ranks, they must acquire the necessary skillset. Where doing so is costly or difficult, only a small share may be able to make the leap, contributing to wage inequalities. Besson also finds that computer use is associated with an increase in the share of the workforce with four or more years of college, even for occupations that do not require college degrees. This may indicate that companies are hiring for higher skillsets rather than training for them.

 Figure 1
Image: James Bessen, Boston University School of Law

The machines are not taking over … yet

In spite of his upbeat message about automation boosting employment and pushing jobs up the value chain, Bessen gives no guarantees that things will stay this way.

“If history is a guide, computers may eventually tend to reduce the number of jobs as more marginal computer applications are exploited that do not produce as much job growth,” he concludes. A case in point: automation in 19th century textile weaving led to growing employment while cloth was in great demand. However, as demand became more saturated and further technical improvements were made, employment plateaued and then declined.

The hope is that, by the time we perceive a similar trend, other technological and wider economic developments will have opened up new job opportunities elsewhere.