Job (and political) Mythology
One of the many myths substituting for facts in the U.S. presidential campaign is that a president can “bring back” manufacturing jobs that have been moved overseas by “un-American” businessmen. Donald Trump is sure it’s so, and even Hillary Clinton seems to flirt with the idea. A similar notion underlies the assertion that Germany “needs” immigrants to replace aging workers.
But, for better or worse, that’s probably wrong.
Rather, we — modern post-industrial societies — seem to be in the early stages of a radical transformation in the relationship between an economy’s need for labor and its capacity to grow, based on an equally radical change in technology’s capacity to execute an ever-widening range of cognitive tasks.
Simply put, “smart” machines are replacing less smart people at a pace that seems likely to accelerate and even more likely to destroy jobs.
In the past, technological progress freed workers from heavy labor and then from repetitive, routine tasks, creating new industries (and new jobs) based on higher productivity and higher knowledge. The process that Schumpeter described as “creative destruction” was at the heart of modern capitalism: lost jobs, ruined companies, and vanishing industries replaced by more productive, dynamic and richer successors. Horse buggies replaced by automobiles; sweat shop laborers by office workers; bank tellers by mobile app writers.
However, something new seems to be happening. Instead of merely replacing manual, routine operations or complementing workers performing cognitive, routine tasks (e.g., secretaries using computers) technology is beginning to demonstrate the capacity to perform non-routine tasks, manual as well as knowledge-based. Obvious examples are automated personal assistants, smart infrastructure (Singapore’s container port is 43% more efficient than America’s), legal research (electronic discovery software is already an integral part of litigation) and — soon — driverless cars (to be followed by trucks and ships).
In the past, workers displaced by technological innovation could be retrained to acquire new skills complementary to or beyond the capacity of machines and computers. Coupled with offshoring — a trend that itself is being undone by technological change — these forces pushed the center of gravity of the U.S. work force from factories to services. Conventional wisdom is that this shift will continue: sure, manufacturing jobs are all but gone (sorry, Donald), but knowledge industries are growing exponentially.
What is actually growing exponentially are increases in computing power and storage capacity as well as the volume and detail of available “big data.” As a result, computerization is increasingly moving well beyond routine, non-cognitive tasks. Robots that have long since replaced skilled workers on assembly lines could soon replace translators, intensive care nurses, and truck drivers, never mind file clerks, waiters, and security traders.
How far can this go? A McKinsey analysis recently concluded that, based on existing technology, 45% of the activities that individuals are paid for today can be automated. While only 5% of workers can be entirely replaced by today’s technology, one-third of the work done by 60% of all workers could be automated without any technological innovation. How that translates into net job losses is unclear, but when McKinsey analysts looked at the insurance industry, they concluded that 35% of existing full time positions are likely to disappear by 2025.
Oxford University’s Carl Frey and Michael Osborne have taken this further. After examining more than 700 different job categories, they concluded that almost 50% of total U.S. employment is in occupations that have a high probability of being automated over the next decade or so, assuming some continued innovation. Indeed, they argue that, in the near term, the bulk of workers in transportation, logistics, office work, administrative support and manufacturing will likely be substituted by computer-driven machines and systems.
Of course, rapid technological change is always disruptive. Queen Elizabeth I famously refused to grant a patent for a knitting machine in the sixteenth century since she feared too many jobs would be lost; the machine eventually got built and hand knitting became a hobby.
Over the long run, technology has demonstrably led to higher productivity, more jobs, and greater social welfare. The unanswerable question is whether we are now approaching a tipping point, where the pace of job destruction could exceed the pace of job creation, at least for advanced societies — particularly for those like the United States that systematically underinvest in education, thereby not only encouraging automation, but also leaving its workforce unprepared for alternative futures.
But if unfettered technological changes really does, on net, destroy jobs, will society try to put the genie back in the bottle? It’s not difficult to imagine a president who — suddenly realizing that most manufacturing jobs aren’t going to Mexico, but are simply going — tries to channel his inner Queen Elizabeth:
“Consider thou what the invention could do to my poor subjects. It would assuredly bring them ruin by depriving them of employment, thus making them beggars.”
Written by Alan Stoga
Originally published May 12, 2016 in EL PAÌS