The Trump Administration has offered a confusing array of justifications for its massive tariff increases—as a negotiation tool to force better trade deals, as a tool to ensure reciprocal tariffs, and as a tool to move more manufacturing to the U.S, especially in industries of national security importance. This post will examine the last justification—the desire to move more manufacturing to the United States. I will address the other justifications in future posts.
In this morning’s Wall Street Journal, James Mackintosh captures well the very different world view that lays behind the Trump tariffs: “I’ve long thought of Trump as focused on people as workers rather than as consumers. The existing system is focused on delivering stuff to satisfy consumer wants, and let jobs fall where they will, even if that is outside the U.S., rather than delivering jobs and supplying only the stuff that ends up being produced.”
It is certainly understandable that many believe that our current consumer focus on trade policy has hurt many Americans. As you can see in the chart below, manufacturing employment in the U.S. plummeted starting in 2000. There are many explanations about this shocking drop, but most economists attribute this to the so-called “China Shock” after China entered the World Trade Organization, which caused a large move of low margin manufacturing to China.
Kyle Handley summarizes the impact:
The China Shock occurred very fast and U.S. labor markets were simply not able to contemporaneously adjust to the jobs lost in manufacturing—which helps explain why, in the absence of robust government retraining programs, the work of Autor, Dorn, and Hanson also finds large increases in unemployment and other social insurance program claims.
The human cost of the China Shock—given that it was concentrated in the Rust Belt—certainly gave Trump’s trade message political resonance and offers some justification for his trade policies.
This China Shock story, however, misses the larger picture. As is apparent from the chart above, total manufacturing employment even in the 1970’s was only 19 million and there were already declines beginning in 1980’s, largely due to both offshoring of low margin operations and technological changes that caused a massive improvement in productivity. Indeed, while manufacturing employment was declining, domestic manufacturing production was increasing. We were making more things with far fewer people.
The bigger picture was that our economy was already moving toward one focused on services (including product design) and even in the China Shock period, job growth in the service sector dwarfed the loss in manufacturing jobs.
Indeed, there is evidence that even within the firms most affected by the China shock, total employment increased, with these firms increasing higher paid positions in “research, design, management and wholesale.” In addition, the manufacturing that moved to China was largely the plants that made standardized goods that required less-skilled labor. As Kyle Handley explains “When the China Shock arrived, the most exposed industries and regions in the U.S. had concentrated in latestage production, i.e. locations with low wages, less innovation, and mature, standardized production processes. This made them less resilient in the face of import competition from China in the 2000s.”
So what does this history tell us about the “move manufacturing to America” justification for the Trump tariffs? It tells us that even the most optimistic upside in employment from the tariffs is limited—we are highly unlikely to see an increase to the 1970 levels because of increases in productivity, but even if we were to reach these levels, the upside is only a few million new jobs—and most likely jobs that are low wage.
The actual number of new manufacturing jobs is likely to be much less. Manufacturing productivity thanks to technology means that the factory floor is much less crowded than it was even as recently as 2000. And the wage deferential for many of these lesser skilled jobs now done overseas is so large that even the massive Trump tariffs might be insufficient to bring these factories back to the U.S. As James Mackintosh argued in his column today, “Bringing back those low-productivity jobs is possible if tariffs are high enough, but will reduce America’s economic lead over the rest of the world. Does America really want to bring back sewing jobs from Bangladesh or Cambodia?”
The potential of four million new manufacturing jobs (in my view wildly optimistic) certainly sounds worth having, but it ignores the larger downsides. We are both workers and consumers, and tariffs will increase the costs we pay for goods and services. The most recent Yale Budget Lab assessment is that each American household will suffer a loss of $4900 in the short term and $2600 in the long term because of the current version of the Trump tariffs. Multiplied over 128.7 million households this is a huge loss: $630 billion in the short term. Even at the most optimistic job levels that means each new manufacturing will be hugely expensive to U.S. consumers (over $150,000 per job).
Moreover, what we gain in manufacturing employment, we more than lose in service jobs. The Yale Budget Lab estimates that the service job losses dwarf the manufacturing job gains, resulting in an increase in the unemployment rate of 0.6 percentage points and a loss of 770,000 jobs.
Why would this be the case?—because the manufacturing that would be brought back to the U.S. by high tariffs are low margin operations, and the cost would be to the high margin operations that we kept in the US. As Andy Kessler explains:
Forget actual trade numbers. Focus on the margin of the products flowing cross-border. Apple has 34% operating margins. Foxconn, which assembles trade-deficit-boosting iPhones, has operating margins of 3%. Which would you prefer?
TVs, cars, clothes, toys and lumber that we import are all low-margin and usually labor-intensive businesses. We export high-margin software, financial services, drugs and AI applications, all intelligence-intensive businesses. I like to say, “we think, they sweat.”
All too often, when people think of service jobs, they think of low paid work in retail and food serve industries. In the modern U.S. economy service jobs more often require high skills and pay very well. But that misses the reality. Service jobs include the engineers that design products. NVIDIA and Apple both outsource almost all of their manufacturing, but do almost all of their engineering and design work in the U.S. by highly skilled and highly paid U.S. workers, whose jobs are classified as service jobs. And almost all of the jobs associate with Artificial intelligence are service jobs—it is a software and data center dominated activity.
Seems to me that the potential upside—even at the most optimistic forecasts—is dwarfed by the huge cost to all of us in our role as consumers and to overall American employment.
Finally, I need to point out that if increasing manufacturing jobs in the US is the goal here, many of the tariffs make absolutely no sense. There are tariffs on goods that—because of climate and geography—we will never make in the United States. There are tariffs on seasonable produce, coffee, and cocoa, which will do nothing to increase jobs in the U.S., but increasing the costs of these items in the U.S. There are also tariffs on the raw materials and components used by U.S. manufacturers. This will make U.S. manufacturing less competitive, and will more than offset any jobs caused by bring production of the components to the U.S. A Federal Reserve Study estimated that higher input costs from Trump’s 2018 tariffs on steel and aluminum actually reduced overall manufacturing jobs.
The China Shock caused real pain that was never adequately addressed, but it does not appear to me that the Trump tariffs are the answer. So what is the answer? Several possibilities come to mind—tax incentives to locate employment in localities hit hard by the China Shock, and generously funded worker retraining programs are options. In addition, given that very practical problems—such as holding mortgages in a declining real estate market—make it difficult for workers to move to where jobs are available, policies and assistance that address these problems can also help. All of these policies cost money, but the cost is much less that the massive costs caused by the tariffs.