Discussions about manufacturing tend to get very contentious. Many economists and commentators believe that there’s nothing inherently special about making things and that efforts to restore U.S. manufacturing to its former glory reek of industrial policy, protectionism, mercantilism and antiquated thinking.
But in their eagerness to guard against the return of these ideas, manufacturing’s detractors often overstate their case. Manufacturing is in bigger trouble than the conventional wisdom would have you believe.
One common assertion is that while manufacturing jobs have declined, output has actually risen. But this piece of conventional wisdom is now outdated. U.S. manufacturing output is almost exactly the same as it was just before the financial crisis of 2008.
In the 1990s, it really was true that manufacturing production was booming even though employment in the sector was falling. During that decade, output rose by almost half. That’s almost a 4% annualized growth rate.
The expansion of the early 2000s, in contrast, saw manufacturing increase by only about 15% peak-to-peak over eight years -- less than a 2% annual growth rate.
And in the eight years between 2008 and 2016, the growth rate has averaged zero.
But even this may overstate U.S. manufacturing’s performance. An alternative measure, called industrial production, shows an outright decrease from a decade ago.
So it isn’t just manufacturing employment and the sector’s share of gross domestic product that are hurting in the U.S. It’s total output. The U.S. doesn’t really make more stuff than it used to.
What’s more, the overall numbers hide serious declines in most areas of manufacturing. A 2013 paper by Susan Houseman, Timothy Bartik and Timothy Sturgeon found that strong growth in computer-related manufacturing obscured a decline in almost all other areas. “In most of manufacturing,” they write, “real GDP growth has been weak or negative and productivity growth modest.”
And, more troubling, the U.S. is now losing computer manufacturing. Houseman et al. show that U.S. computer production began to fall during the Great Recession. In semiconductors, output has grown slightly, but has been far outpaced by most East Asian countries. Meanwhile, trade deficits in these areas have been climbing.
In other words, Asia is still solidifying its place as the workshop of the world, while the U.S. de-industrializes. The 1990s provided a brief respite from this trend, as new industries arose to replace the ones that had been lost. But the years since the turn of the century have reversed this short renaissance, and manufacturing is once more migrating overseas.
Manufacturing skeptics often draw parallels to what happened to agriculture in the Industrial Revolution. But the two situations aren’t analogous. In the 20th century, U.S. agricultural output soared even as it shed jobs and shrank as a percent of GDP. Machines replaced most human farmers, but the total value of U.S. crops kept climbing.
Meanwhile, the U.S. to this day runs a trade surplus in agriculture even as it runs a huge deficit in manufactured products. America pays for computers and cars and phones with soybeans and corn and beef.
So U.S. manufacturing is hurting in ways that U.S. agriculture never did. The common refrain that the modern shift to services parallels the earlier shift to industry might turn out to be true, but the parallels are not encouraging.
Faced with this evidence, many skeptics will question why the sector is important at all. Why should a country specialize in making things, when it can instead specialize in designing, marketing and financing the making of things?
This is a legitimate question, but there are reasons to think a successful developed nation still needs a healthy manufacturing sector. Harvard University's Kennedy School of Government economist Ricardo Hausmann believes that a country’s economic development depends crucially on where it lies in the so-called product space.
If a country makes complex products that are linked to many other industries -- such as computers, cars and chemicals -- it will be rich. But if it makes simple products that don’t have much of a supply chain -- soybeans or oil -- it will stay poor.
In the past, the U.S. was very successful at positioning itself at the top of the global value chain. But with manufacturing’s decline, the rise of finance, real estate and other orphaned service industries may not be enough to keep the country rich in the long run.
More top economists are starting to come around to the view that manufacturing is important. Massachusetts Institute of Technology economist David Autor, in a recent phone conversation, told me he now believes that the U.S. should focus more on industrial policy designed to keep cutting-edge manufacturing industries in the country.
He cites Sematech, a government-led consortium that tried to help the U.S. retain its lead in semiconductor manufacturing in the 1980s and 1990s, as a successful example of high-tech industrial policy. The stellar performance of semiconductor manufacturing in the 1990s and 2000s relative to other industries in the sector, as reported by Houseman et al., seems like something the U.S. should aim to emulate with next-generation industries.
So U.S. leaders should listen to manufacturing skeptics a little bit less, and pay more attention to those who say the sector is crucial. It's worth noting that President Donald Trump, who was elected on a promise to restore American manufacturing, has shown more interest in cutting government programs designed to give industry a helping hand. If there’s going to be a U.S. industrial policy renaissance, it might not be his administration that leads it.
By Noah Smith