[Analytics] Automation hypochondria in the new industrial age

A robot arm makes an ice cream for customers at a fastfood restaurant of KFC in Shanghai, China, 25 August 2019 (Photo: Reuters). Sketched by the Pan Pacific Agency.

The world has experienced four industrial revolutions and as each has unfolded, so have dire predictions of massive job losses. Looking back at the first three, it is clear that concerns were misplaced. The number of jobs increased each time, as did living standards and every other social indicator. McKinsey predicts that 800 million workers could be displaced in 42 countries — a third of the workforce — because of the Fourth Industrial Revolution (4IR). Although similar predictions were made at the onset of each revolution of the past, could there be something more to it this time? Jayant Menon specially for the East Asia Forum.

Disruptive technologies — artificial intelligence, robotics, blockchain and 3D printing — are indeed transforming social, economic and political systems, often in unpredictable ways. The technology itself is difficult to map because its growth rate could be exponential, factorial or higher. It is this unpredictability that is making impact assessments difficult. But not impossible.

Many low-skilled, repetitive jobs are being automated, starting in high wage countries but quickly spreading to the developing world. With two-thirds of the world’s robots already in Asia, some expect this region to be particularly susceptible to these changes.

But are there limits to automation? To answer this question, it is essential to first understand how work is transforming, especially within global value chains (GVCs). GVCs dominate production in the exportable sector of most Asian countries. Jobs associated with GVCs consist of a bundle of tasks, true at all skill levels. As long as at least one of the multitude of tasks that a worker performs cannot be technically and economically automated, then that job is probably safe. There are lots of jobs like that within and outside GVCs, even though it may not appear so on the surface.

For example, while most tasks performed by waiters can be automated, human interaction is still required. Human hands are also highly complex and scientists have yet to replicate the tactile sensors of animal skin. The robot may deliver your soup, but struggle to place it on your table without spilling it.

The debate also tends to wrongly focus on gross rather than net jobs, usually unintentionally. But it is the net figure that matters in this debate.

For instance, greater automation of production processes will require greater supervision and quality control. Humans will be required to carry out this function. The focus on gross job figures ignores the higher skilled jobs created directly as a result of greater automation.

As long as the cost of adding more supervisors does not outweigh the savings from automation, the reduction in the price of the final good would spur an increase in demand. If the increase in demand is large enough, it may even expand the number of jobs in factories that automate part, but not all, of their production process. In this case, automation leads to a net increase in jobs.

There will also be inter-industry effects. Productivity gains from new technology in one industry can lower production costs in others through input–output linkages, contributing to increased demand and employment across industries. Higher demand and more production in one industry raises demand for other industries, and on it goes.

Why then the widespread pessimism about the 4IR and jobs?

It could be that it is easier to see how existing jobs may be lost to automation than it is to imagine how new ones may emerge sometime in the future. In a sense, this is like the gross versus net confusion, but separated by time and greater uncertainty.

It is also more sensational to highlight the job displacing possibilities than the job creating ones. Since the effects will vary not just between developed and developing countries, but also within each group, we may hear more about countries that suffer net job losses now or soon, which then shapes our overall negative perception. We may also hear more about it because while the benefits are widely dispersed across the general public through lower prices, the costs are concentrated and can displace low-skilled workers, providing greater incentive to organise and lobby against or complain about the negative effects.

When there is enough uncertainty, it is generally safer to overstate rather than understate the potential cost to innocent victims of change. All of these factors combine to explain the unwarranted pessimism over jobs.

But there could be a silver lining. If negativity leads to greater efforts to reskill and reshape the workforce to better adapt to change, exactly what is required, then there is no overdoing it. Ironically, it could well be this pessimism that produces the preparedness that results in it being misplaced. But because the effects will vary across countries, so too must the required preparation. It is not only the extent of reskilling that will vary, but also the type, depending on a host of factors which broadly correlate with the level of income.

Jayant Menon is a Lead Economist (Trade and Regional Cooperation) in the Office of the Chief Economist at the Asian Development Bank (ADB), Manila.

Share it

Exclusive: Beyond the Covid-19 world's coverage