The world appears to be enjoying its broadest economic expansion in more than a decade, and about 80% to 90% of the global economy is considered to be at full employment. Do we really have to worry about robots replacing workers?
Recent published studies from the estimable environs of Oxford University and the Massachusetts Institute of Technology claim that by 2023 automation will have taken over 47% of the current jobs in the U.S., 69% in India, and 77% in China. Others think that may be a bit of an overreach.
Using Department of Labor statistics, the economic research team at Putnam Investments makes the historical observation that “jobs don’t get automated away, tasks do.” They say today’s typical job involves more tasks than ever. Across some 18,000 work activities, they identify 41 discreet tasks categorized as either basic, repetitive, or advanced.
Human progress is a long-running saga of automation and organization taking on those basic and repetitive tasks. The process drives jobs toward a richer mix of advanced activities, including designing, refining, strategizing, and managing client and co-worker relationships.
Nevertheless, artificial intelligence (AI) is advancing into areas we have associated with distinctly human capabilities. Economists at BCA Research recently went looking for hard evidence that the displacement effect of new technologies is any stronger than in the past.
They do confirm that robot capabilities have improved significantly, while quality adjusted costs have plummeted. According to the International Federation of Robotics (IFR), robots used in industry are approaching two million units, having grown about 10% per year since 2010. Automakers represent 45% of that total with another 17% deployed in the computer and electronics sectors. Most of the balance is used in producing metals, chemicals, and electrical/electronic appliances.
Japan was an early leader in deploying robotic technology, consistent with its prominence in autos and electronics and slow growth in its working age population. However, Japan’s stock of robots has actually fallen off the past 20 years. Over that same period, China saw the sharpest increase in the sheer number of industrial robots, and U.S. deployment has climbed steadily.
A more telling measure is a country’s robot density: the number of such devices per 10,000 people employed in manufacturing. By that gauge, the clear leader is South Korea with over 600 robots per 10,000 manufacturing workers, followed by Germany and Japan above 300, and the U.S. with about 200. China’s current robot density at about 75 is just over the average of the 23 countries in the IFR’s database.
That still leaves the question of whether robots’ expanding capabilities will drive greater worker displacement and a broader deflation effect. If so, we would expect to see rising unemployment and productivity (output per hour worked) and stagnating real wages. But so far, the rise of robots has been accompanied by a recovery in employment, rather weak productivity gains, and mostly recently, real wage gains.
News reports that show robots performing human-like actions on the factory floor may heighten our sense of impending job displacement. But the folks at BCA Research see robots as “just another chapter in a long history of automation.” Precise comparisons of major technological advances are difficult, but the Harvard Business Review has estimated that robots contributed about 0.36% of the 2.5% annualized productivity growth from 1993 through 2007, compared to the steam engine’s 0.34% annualized contribution from 1850 to 1910, and 0.60% from the information technology revolution from 1995 to 2005.
Using data from the U.S. Department of Labor, the IFR has been able to show modest correlation across different industrial sectors between higher robot density and lower unit labor costs. That may simply be trailing confirmation of the industries in which current robot technology has been most applicable. As for a broad deflationary effect, a comparison of inflation in major industrialized nations has not yet shown a correlation with robot density.
Integrated information technology, robotics, and artificial intelligence will continue to assume tasks traditionally done by humans while reshaping and often enriching the jobs we do. It may be foolhardy to claim that any job is immune to being “automated away.” But it is estimated that 70% of today’s occupations require creative thinking, still considered a distinctly human capability, and in the near future more than 80% of jobs will revolve around advanced tasks. Progress certainly can be unsettling, but probably less so than the alternative.