A Chinese factory made international headlines two years ago by firing 90 percent of its workers, while simultaneously making 162.5 percent more stuff. How did they do it? Automation.
Since then the mainstream media has worked itself, and many Americans, into a frenzy over “job-stealing” robots. Automation will cause mass unemployment and poverty, they say. Therefore, we should tax the robots, and use the proceeds to fund a universal basic income—presumably administered according to the maxim from each according to his ability, to each according to his needs, or something like that. Even otherwise pro-business, economically literate publications like Business Insider have jumped on the bandwagon.
All this raises the question: should Americans worry about robots taking their jobs? More specifically: will automation cause mass unemployment?
No. And no.
The “problem” of automation is nothing but sensational clickbait—fear sells.
In reality, automation does not impact long term employment rates. In fact, automation, via technological growth, is the only way to grow the economy in the long run. Technology makes us rich. And this is not a new observation—back in 1964, when America was having this exact same debate, the same arguments were presented. Some things never change. This is the bottom line: the robots didn’t take our jobs then, and they won’t take them now.
How Employment Works
At a bare-bones level, employment is determined by the ratio between productivity and output. All other things being equal: higher productivity (getting more done each hour) means fewer jobs, higher output means more jobs—if both productivity and output increase at the same rate, then employment is not affected.
Imagine, for example, an all-American car company called Aspen Automobiles. Aspen makes 1,000 vehicles every year at its factory, which employs 100 workers. In 2016 the automaker got lucky, selling out of its popular Poplar Crossover. Given the company’s success, Aspen has big plans for 2017, with 100 extra Poplar Crossovers expected to roll off the assembly line. This means Aspen must hire 10 more workers, since 10 percent more output requires 10 percent more labor. This is how increasing output—also known as economic growth—creates jobs.
Now let us add another factor into the equation. Pretend 2016 was a normal year, and Aspen does not think it could sell more vehicles in 2017. Nevertheless, Aspen is determined to make more money. Therefore, the company invests in a few robotic welding torches that allow line workers to weld much faster than they did before. This improves the factory’s efficiency by 10 percent. Now Aspen can make the same number of vehicles with only 90 employees. In this case, higher productivity via automation cost workers their jobs.
Now combine the two factors: 2016 was a great year, and Aspen decides to go ahead and make 100 extra vehicles. Not only that, the company also invests in the robotic welding torches, making the factory 10 percent more efficient. What happens to the workers? On the one hand, Aspen needs more employees to make more vehicles; on the other, the manufacturer needs fewer workers because of automation. Overall, employment is not affected in a meaningful way because the two factors cancel each other out: Aspen could make more vehicles with the same number of employees. This is what makes an economy more prosperous.
The takeaway: if productivity and output increase (or decrease) at the same rate, employment does not change.
New Technology, Old Fears
The real economy works the same way. There are countless historical examples that make this point, but none are better than the Industrial Revolution. Between 1780 and 1820 Britain’s economy changed rapidly: steam-powered machinery replaced human workers at an alarming pace—so alarming, in fact, that pogroms of workers literally waged war against industrial equipment. This is where the term Luddite comes from. As it turns out, today’s automation is nothing compared to what happened back then—the power loom alone made British textile weavers 40 times as efficient. Yet despite this, Britain actually had a severe labor shortage because output grew faster than productivity. The end result was that Britain got rich.
Despite what you have been led to believe, this still holds true today, and the proof is in the data. Looking specifically at the manufacturing industry: between 1950 and 1979, manufacturing employment increased because output grew faster than productivity. This changed over the next decade, however. By the 1990s, the historic balance was upset. Between 1989 and 2000 American manufacturing output grew by 3.7 percent on average, while productivity grew by 4.1 percent—employment consequently declined. But since 2000, output growth nosedived: output grew only 0.4 percent per year, on average, while productivity increased at a rate of 3.7 percent. As a result, America shed more than 4 million manufacturing jobs.
What did the media blame for America’s job loss? Automation. Yet the historical data refutes this claim unambiguously: automation does not cause job loss unless output growth lags behind. So the real question is: what is causing output growth to decline?
The answer is simple, and it is one of the main reasons Donald Trump won in 2016: the trade deficit.
The fact is that much of America’s new output growth is occurring abroad, as opposed to domestically—rather than build a new factory in Michigan, we build it in Mexico. Rather than open a call center in Philadelphia, we set it up in the Philippines. We consume more and more goods and services, but do not make them ourselves. As a result, output stops growing, but productivity does not. This gives the false illusion that, by increasing productivity, automation is causing job loss—but this is only the proximate cause.
The deeper issue is output growth stagnation caused by offshoring, and reflected in America’s balance of trade.
The bottom line: stop blaming robots. Blame China’s predatory trade policies, or our politician’s unwillingness to address them. We can never solve the problem if we spend our time jousting windmills.