Automation and offshoring are no impediment to building things in America.
Automation and offshoring are no impediment to building things in America.
P erhaps no economic idea has penetrated the American mainstream more successfully than “the manufacturing jobs aren’t coming back.” It’s a high-status position to hold, shared by public intellectuals of all stripes, from New York Times editors to Ivy League economists to American presidents, trickling down to a thousand armchair policy wonks. Prestigious publications churn out article after article repeating this position, and the numbers appear to back it up: American manufacturing employment has been declining for 30 years, even as we produce more manufactured goods than ever before.
The conventional wisdom goes something like this: American manufacturing employment will continue to decline, because automation has reduced the need for human factory labor and globalization means that manufacturers will offshore to cheaper countries whatever labor is left. The jobs of the future will instead come from the service sector, which is more innovative and dynamic. Instead of promising the return of manufacturing jobs, our leaders should invest in education to nurture knowledge-economy skills and introduce some version of Universal Basic Income for anyone who gets left behind.
Those who have absorbed the conventional wisdom might ask a simple question: If American manufacturing employment is declining, should we even care?
Yes — because manufacturing is special. The sector has certain unique, desirable qualities that other sectors lack. The first is that a healthy American manufacturing sector keeps our nation secure. The COVID-19 pandemic highlighted the importance of domestic supply chains when our economy ground to a halt for want of capacity to produce ventilators, protective equipment, pharmaceuticals, and other essential goods. Instead, our lack of manufacturing capacity made us dependent on China, our chief adversary.
This dynamic is replicated in our defense supply chains. A 2018 Department of Defense report revealed that “dozens” of militarily significant inputs had only one or two domestic suppliers — and that we depend on China to manufacture many of the products used to build our ships and armaments. As the number of Americans working to produce these critical materials shrinks, the know-how necessary to survive the next global crisis shrinks with it.
Manufacturing has special growth advantages. Manufactured goods are tradable, meaning that they can potentially be sold anywhere in the world. On the other hand, most service-sector products are constrained by geography — your local barbershop or cable provider won’t be selling their services in India. Manufacturing has a job-multiplier effect that creates four to five additional jobs for each manufacturing job, while the service sector multiplier is only 1.5.
Manufacturing also has a growth advantage through its ability to absorb increasing returns. Because manufacturing is susceptible to R & D innovation, output can rise faster than a given increase in input, allowing the manufacturing sector to gainfully absorb rising capital investment by creating economies of scale.
Simply put, not all industries are created equal: For example, having a domestic auto-manufacturing sector creates demand for a large supply chain of high-value components such as electronics, chemicals, and steel. Not so for a car-rental business in the service sector.
Manufacturing has an additional, especially important advantage: It can employ large numbers of comparatively low-skilled, low-education workers to do high-productivity work, meaning high wages. The average manufacturing employee makes 13 percent more in hourly wages than a comparable employee elsewhere in the private sector, and the reason for this wage premium is simple: People don’t pay very much for the products of less-educated workers in the service sector. While highly educated service-sector workers such as doctors and consultants can command high salaries for their work, a less-educated service-sector worker might be cutting hair or flipping burgers — and people will pay only so much for haircuts and burgers. If you don’t have a degree, your best chance at a high wage will be producing higher-value goods such as cars or computers. The manufacturing sector’s special ability to place less-educated workers in a position to make high-value goods is a unique social benefit.
This thinking tends to make people uncomfortable. We’d like to imagine that everyone has the potential to become a doctor and or a consultant, so we design our education system to push everyone to the heights of the service economy and leave vocational education as an afterthought. Teaching people to be better steelworkers feels like admitting defeat; we’d rather teach steelworkers to code. Unfortunately, we have no idea how to do that. Worker-retraining programs have been dismal failures, and while college enrollment continues to rise, actual B.A. attainment has remained largely flat.
Meanwhile, the premise behind an education-focused strategy — that we can do high-value knowledge-economy work here in America while our innovations get manufactured elsewhere — is flat wrong. “Innovate here, build there” doesn’t work because firms reap efficiency gains from locating engineering, design, and R & D next to the factory floor. Increasingly, the innovation is migrating to countries such as China, which has gobbled up an enormous share of the R & D spending that used to take place in America. An innovation economy without a strong manufacturing base is a fantasy. Education and retraining for the knowledge economy isn’t so much a plan as the desperate refuge of policymakers without a plan.
This planning failure characterizes much of our country’s experience with the transition to a service economy: For tens of millions of people, what replaced manufacturing isn’t working. When the promised panacea of education and retraining failed, it disrupted the core assumption of the neoliberal economists who predicted a painless transition: that workers would simply move into more efficient industries when they lost their old manufacturing jobs. Instead, as the economist David Autor has shown, manufacturing workers tended to stay put and make less (or exit the workforce entirely) rather than successfully relocate and retrain.
The result was a nasty collection of “negative social externalities”: falling marriage rates, more single parenthood, more child poverty, a worsening opioid crisis, and the deaths of despair famously chronicled by Anne Case and Angus Deaton. The economists’ spreadsheets might have said that these workers were compensated for their lost earnings by lower consumer prices, but an unprecedented decline in life expectancy for middle-class Americans suggests that the economists were missing something.
The “American carnage” for blue-collar workers in the post-manufacturing economy proves that the decline of the manufacturing sector is nothing to celebrate. But for many, this doesn’t matter, because automation and offshoring will sap American jobs from the factory floor regardless, even as the remaining factories are highly productive. Many economists and policymakers think that these processes are inevitable, and so, while celebration is inappropriate, we must resign ourselves to our fate.
But this course of affairs is not inevitable. Both automation and offshoring have been fundamentally misunderstood by a large section of the commentariat. In reality, despite the pundits’ resignation, the American manufacturing sector can meaningfully increase its share of employment.
Automation is the name we give for the decline in man-hours necessary to produce the same unit of output. As workers become more productive with the aid of technology, it takes fewer of them to produce the same amount of stuff. This process has been ongoing since the dawn of the industrial revolution, but it didn’t eliminate jobs because it also allowed for massive increases in output. In the 1950s, American manufacturing output increased at around 3.5 percent a year, leading to employment of about 17 million in manufacturing throughout that decade.
But by 2000, output growth had slowed to a mere 1.1 percent per year. Meanwhile, worker productivity growth also slowed — the opposite of what you’d expect if automation were eliminating jobs on its own — leading to a drop in manufacturing employment to 12 million as fewer workers were needed to produce comparatively less stuff. As Oren Cass noted, if output had instead increased at a constant rate from 1950, manufacturing employment would have actually increased to 18 million.
The impact of automation on manufacturing has been exaggerated by misleading measurements of output and productivity. When a third of manufacturing employees lost their job between 2000 and 2016, economists advised that the sector was still strong because output and productivity were still growing. The lost jobs were the result of inevitable automation, they concluded. But economist Susan Houseman made a discovery: Output and productivity looked healthy only because virtually all of their growth came from mathematical “quality adjustments” to the computing industry — specifically, the increased speed of processors was counted as additional output, even though no more products were actually being shipped. By controlling for this adjustment, Houseman revealed that manufacturing output was actually flat, growing at a mere 0.2 percent per year, and had even declined between 2007 and 2016. Houseman’s conclusion: Automation was not responsible for the sharp decline in manufacturing employment. The real problem is that American factories are producing fewer goods.
Nothing was inevitable about the output slowdown. While it is true that the average labor content of manufactured goods is falling, the amount of potential output expansion is enormous. As people grow richer, they consume more manufactured goods — and the ceiling for consumption has barely been scratched. This is strikingly illustrated by Peter Menzel’s photographic project Material World, in which he asked average-income families in different countries to pose in front of their homes with all of their material possessions assembled outside. Families from the developed world show off cars and expensive appliances, while developing-world families own far less — sometimes little more than farming implements, utensils, and a few items of furniture. As these countries grow richer, it will entail large increases in consumption of manufactured goods, and those goods will get made somewhere.
So if someone needs to sell cars and computers to the rising African middle class, who will? Americans imagine that the people doing their old manufacturing jobs are low-wage workers in developing countries such as China and Mexico. In fact, American manufacturing has been out-competed by other developed countries. While only 8 percent of American employees work in manufacturing, the rate is 17 percent in Japan and South Korea, 18.5 percent in Italy, and 19 percent in Germany. These are fellow First World nations with no necessary cost advantage over the U.S., and yet they’ve managed to make manufacturing a far larger piece of their national economies. While manufacturing makes up only 12 percent of America’s economic output, it’s nearly a quarter in Germany, the economic leader of Europe. These countries employ more people despite embracing automation — they use far more robots in their factories than the U.S. does.
Even in supposedly lower-cost countries such as China, the story is not so simple. The majority of China’s cost advantage comes from factors that have nothing to do with labor costs, such as government subsidies, which are responsible for an estimated 16.7 percent of China’s overall cost advantage. While low wages are largely a function of a country’s development stage and are therefore difficult to surmount through competition, policy-driven factors — such as tax policy, subsidies, infrastructure investment, currency manipulation, and environmental regulation — are major determinants of nations’ ability to compete for global manufacturing share. This is where the U.S. is losing. The average cost disadvantage between American manufacturing and its lower-wage competitors is estimated at only 17 percent — well within the reach of smart policies, if only we could enact them.
We should start by rethinking trade policy: what the Senate Finance Committee once described as “the orphan of U.S. foreign policy.” For a long time, the dominant view was that trade deficits didn’t matter — if other countries want to subsidize their exports with taxpayer money so they become cheaper for American consumers, that sounds like a free lunch. But if that keeps happening in every industry, American industry will find it impossible to compete, and suddenly America won’t have anything to trade for those imported goods — except debt and assets. That’s what a trade deficit is: mortgaging your future productive capacity so that you can consume more than you currently produce.
Trade policy should therefore start from the premise that unreciprocated free trade — an open U.S. market for foreign imports but closed foreign markets for U.S. exports — is the worst-case scenario over the long term. These are the conditions that cause American producers to lose market share. Our trade policy should identify strategic industries — sectors such as robotics, aerospace, telecom, and others, along with industries important for national security — and do everything we can to win market access for them abroad while keeping American producers healthy at home.
This might sound elementary, but that’s not how American trade policy has worked. Instead, we historically gave away market access, believing that low consumer prices were their own reward. Economist Ian Fletcher summarized the problem in his 2011 book Free Trade Doesn’t Work: “Having disarmed ourselves by throwing open our markets, we desperately need to disarm everyone else by forcing their markets open too. But we try to do this after having thrown away our principal leverage: access to our own market.” From 1820 to the dawn of World War II, America built its industrial juggernaut with average weighted tariffs of 20 percent. Today, they’re 2.85 percent.
Clyde Prestowitz, a trade official in the Reagan administration, recounted in his 2010 book The Betrayal of American Prosperity how American trade policy would prioritize low consumer prices above all else. In the 1980s, American firms that manufactured hydro-generators — the enormous machines that turn water power into electricity — came under pricing pressure from new Japanese entrants into the market. The Japanese price advantage wasn’t due to a better product, it was due to a better industrial policy that exploited the rules of the market for hydro-generators. The Japanese firms developed a strategy to take turns submitting the low bid to win manufacturing contracts — on one bid Toshiba would come in far below the pack, the next time Toshiba would join the rest and Hitachi would bid far below. They repeated this until the American firms were forced to either offshore production or exit the market.
Siemens-Allis, the last American manufacturer, appealed to Prestowitz for help from the government, which could have brought a trade suit against the Japanese for selling products below cost. But Reagan’s Council of Economic Advisers didn’t want to act. Prestowitz recalls what happened:
One council member told me he didn’t see the problem. As far as he could tell, American electric companies and municipalities were getting inexpensive generators, and consumers were thus presumably benefiting from lower electric rates. Nor was he worried about the fate of Siemens-Allis workers who would lose jobs if the company closed operations. They could easily find something else to do in the more robust services economy.
Siemens-Allis laid off its workforce, who presumably found the much-vaunted opportunities in the service sector lacking.
When our trade policy isn’t following the wrong strategy, it frequently fails to follow any strategy at all. Our trade priorities have too often been influenced by companies interested in offshoring production instead of the pursuit of an overarching national strategy that benefits American manufacturing. This occurs through the mechanism of Industry Trade Advisory Committees, which consist of private-sector businesses that tell the government what we should seek to gain in our trade agreements.
Some might protest that the very idea of a national trade strategy smacks of planning — the dreaded “picking winners and losers” that free-trade doctrine is designed to avoid. But having no strategy doesn’t mean the market will pick winners, it means that Chinese policy will. Free trade with China means allowing their prices — distorted by subsidies and currency manipulation — to shape our own market, where American firms will struggle to compete without the benefit of an industrial policy of our own. As Marco Rubio put it in a recent report on China’s rise: “In a world of state competition for valuable industries, a domestic policy of neutrality is itself a selection of priority.”
Once we grant that manufacturing is special and that market fundamentalism has weakened it, a whole world of policy options long deployed by our competitors suddenly becomes attractive. It seems as if we’re slowly waking up. Robert Lighthizer is the first U.S. trade representative in recent memory to explicitly prioritize shrinking the trade deficit. There is bipartisan agreement that the government needs to do more to protect strategic American industries, resulting in ambitious proposals such as the American Foundries Act of 2020, which would authorize $15 billion in funding to incentivize manufacturing of critical semiconductors at home. In a hopeful sign, the bill originated in Senator Tom Cotton’s office and was later joined by Chuck Schumer and other co-sponsors from both parties. There’s more to do — such as establishing a National Manufacturing Foundation that would unite a series of piecemeal manufacturing R & D programs, ensuring that American innovations are commercialized and scaled in the U.S. for the benefit of American workers.
Now that automation and offshoring are no longer credible culprits for the historic collapse in American manufacturing employment, policymakers have no more excuses. This decline isn’t inevitable — it’s a choice. The next time you hear that “those jobs aren’t coming back,” hear it for what it is: an abdication of our duty to the millions of Americans who have watched their middle-class aspirations slip away.
These views are the author’s alone.
This article has been emended since its original publication.