The Trump Tariffs: A Great Inefficiency Machine

His tariffs will neither generate the desired revenues nor revitalize American manufacturing

Really, another substack on tariffs?

Yes. There will be plenty of articles on tariffs, but I taught trade to Econ 101 students for decades, and believe that experience lets me convey the key issues clearly and succinctly.

Tariffs as an inefficient tax

First, tariffs are a tax on consumers: importers pay them and pass them on to their customers. They are not paid by the exporters. That makes them an inefficient way to raise revenue, since they drive up prices of domestic products while depressing the quantity of competing imports. We as consumers thus pay higher prices on goods of all sorts, whether Made in America or imported. But while we will consume less overall, it’s the quantity of imports that falls the most (or for many goods, goes to zero). The government only collects money on that shrinking slice. The government won't collect much revenue, certainly nowhere near the $600 billion amount touted by Peter Navarro, an economist specialized on the US electric utility industry who somehow became Trump's favored advisor on international trade and China.

Tariffs focus on manufacturing; the US is a service economy

Second, tariffs ignore the reality of the US economy. The core insight goes back to Adam Smith writing in 1776 about the benefits, individually and collectively, of expanding the market to enable specialization. Think of the NCAA basketball playoffs. They're interesting because of the skill and athleticism of the players, and also because basketball makes for really good TV. For the "students" who are playing, it's a full-time job. They don't sew their own uniforms, bake their own bread, or even spend much time hitting the books. Instead, they trade for those. They do not try to be the jack of all trades, they aim to be the master of one. We like it that way, even if sadly most college athletes are unable to turn their mastery into post-graduation income.

David Ricardo extended Adam Smith's insight with his concept of comparative advantage, that countries benefit from specialization just as individuals do. Let's think of the garment industry and a modern farmer.

Operating a sewing machine takes dexterity and the ability to do the same task for hours on end, day after day, so it's not for everyone. However, that's a skill set that most people can handle. It's also labor intensive. The array of specialized sewing machines used today are the same ones used since the late 1800s. Sure, in 1900 the reliance on a large steam steam engine meant that sewing machines had to be heavy enough to be stable when powered by a belt running from an overhead shaft. Today's are small, portable machines powered by an electric motor. That has lowered the cost of setting up a garment operation, but has done little to improve the productivity of the workers. Sure, there are lots of little advances, computer-driven water jets to cut cloth to pattern instead of using scissors. Sewing on a belt loop to trousers and skirts, however, remains a hand operation. The labor input remains large, and the investment in equipment modest.

Not so with modern farming. The two brothers who ran the dairy farm a couple miles from me carried calculators in their pockets, and had master's degrees. They had to know how to care for sick cows, they were trained in artificial insemination and in using the databases on the genetics of the bulls whose semen they bought so that they could improve the characteristics of their milk cows. They had to know how to repair machinery, to weld and to machine parts because they couldn't afford to wait for a repairman – cows have to be milked twice a day, whether your milking machine is operating or not. They had to be proficient in growing corn and alfalfa, from soil chemistry to which hybrids matched their climate to what dietary supplements they then needed given the nutritional value of their current variety of corn. How much fertilizer to apply? – that's why they carried calculators, because that was a function of how much fertilizer cost, the fertilizer responsiveness of the corn they had planted that year, which was poorer when the weather was dry, and the price of milk. They had to maintain a database on each cow, to monitor how much milk it produced, milkfat levels, how to adjust feed, and when to "freshen" the cow by inseminating her to bear another calf. Oh, and then what to do with the boys, er, steers. If the corn harvest was good, they'd keep them longer, as a larger steer got a better price at auction. It's daunting to acquire the requisite skill set, and requires flexibility and a high degree of physicality. Few people have what it takes to be a modern farmer.

So the outcome is that Bangladesh benefits by exporting garments, and importing food (though probably not much dairy). Yes, Bangladesh remains wedded to agriculture, but their farmers don't have the years of education and the tractors, milking machines, refrigerators, and the like to make them productive. In contrast, we in the US thankfully don't have the reserve army of unemployed available in the tens of thousands to sew garments. Plus our healthcare sector is 40% bigger than all of manufacturing. If tariffs were focused on promoting just one industry, OK. That’s not what the President has done. We aren’t going to shut down hospitals so we can increase the domestic production of undergarments.¹ Trade is a great labor saving machine, and as a service economy we can’t afford to have a major slice of our population dedicated to producing bras and briefs.

A historical aside

Reforms after WWII facilitated such trade, even if the identity of our trade partners has shifted while at the same time the capital- and skill-intensity of US agriculture became qualitatively higher. Trade revolutionized both industries over the last 75 years, as the GATT (starting with 23 countries in 1947) and the MFA [Multi Fibre Arrangement] of 1974 lowered trade barriers in textiles. The WTO superseded the MFA and extended coverage to a wider set of countries, including China from 2001. Lower trade barriers likewise expanded the market for agricultural products – trade has been a group dance, not until today (as in April 3) a solo enterprise with the lights lowered...

Trade technology also changed. A lot of the costs we see in a clothing store come from the cost of sourcing goods, shipping them to stores, and doing so to provide an inventory that is both seasonal and stylish, and in an appropriate array of sizes. Additional costs came from unpacking and hanging garments on racks, assisting customers in finding a match in size and style, and then handling purchases. Inventories also had to managed, marked down as the end of the season approached, with the remainder then taken to a landfill. There was a lot of waste, and that inefficiency kept prices high. The over the last 5 decades the deregulation of truck transport, the development of containers and specialized ships, and finally computers and then the internet lowered those costs. Today a factory in Vietnam or Bangladesh might sew clothes only when an Amazon order was in hand, delivering only the desired color and size and thereby eliminating the costs of inventory, particularly unsold inventory, and other expenses attendant to bricks-and-mortar operations. Overhead was reduced, too, as buyers for a store no longer had to travel to fashion shows, or fashion houses travel to remote factories to coordinate production.²

Even our language changed to reflect greater trade

The shifting sense of the term "wardrobe" is indicative of that transition. It used to refer to a piece of furniture in which clothing was stored; most 19th century houses didn't come with closets. Indeed, the first house of my brother-in-law in the Appalachian foothills of Virginia had none, and it probably dated from the 1930s. I grew up in a house in Detroit built in 1930; it had small closets. At the time, though, Detroit was one of the wealthiest cities in the US. Then there’s my current house, 2010 construction but initially with one bedroom, no mini-mansion for my wife and I. It does, though, have a large his-and-hers walk-in closet, and it’s full.

Tariffs will mean empty rather than overflowing closets. That's because we aren't going to return to the era of Southern textile towns, which in turn was a successor to New England textile towns – I went to grad school in one such, New Haven, Connecticut, which in 1980 bore all the marks of industrial glory followed by industrial decline. As I mentioned above, we’re fortunate in the US to have begun the Trump administration with low unemployment, but people at the low end of the wage spectrum already have a hard time making ends meet. Young couples don't earn enough to afford child care, unless they have access to a non-profit one with a sliding scale. So one member of the household has to stay at home, unless they have willing and able grandparents nearby.³ So we won't see sweatshops returning to the rural South.⁴ We'll simply be paying a lot more for our clothing, and buying a lot less of it.

Other sources of specialization and trade: industrial districts

That's the comparative advantage story. But there's also the "new" trade theory of specialization that focuses on the role of industrial districts and on industries with many similar products, such as automotive, that are made in multiple factories.⁵ As it happens, industrial districts matter for garments – on a trans-Pacific flight I once sat next to the chief purchaser for a university apparel supplier for schools up and down the East Coast. In the early days he had headache after headache in China. Universities are finicky about school colors, and matching colors isn't easy. Using the same dye, this year's cotton may not come out the same color as last year's. Chinese aesthetics expect different characters in a poem or logo to be uneven in font, mixing large and bold with smaller and thinner ones. Spacing isn't uniform. Advertising slogans might not be centered or horizontal. Unfortunately, that just won't fly on an expensive sports jersey.⁶ So he had to spend a lot of time working with his suppliers in their factories to instill in them the expectations of the university stores stocking his wares. That changed over time, as there gradually arose a whole network of specialized firms that focused on color matching, on logo design and placement, and on folding, packaging and shipping garments. There were dye suppliers, sewing machine vendors, used equipment brokers, button and hook and frill and lace vendors. Those early (and problematic) factories grew into an industrial district.

Industries where such agglomerations develop become geographically sticky. Moving one plant doesn't work, because the productive ones are able to specialize because they can rely on a local network of suppliers, design specialists and repair services. Take those away, and the plant incurs a lot of extra costs, potentially flying someone in from overseas at great expense to spend 2 hours repairing a specialized piece of equipment, or solving a color matching problem. An example that's important to the auto industry are plastic injection molds. There is no longer a pool of skilled machinists adequate to supply the needs of a car company, in part because demand peaks when a car company is prepping to launch a new model, and that order flow is intermittent. Given other changes in the industry, autos alone can’t support a group of firms or provide career prospects good enough to encourage young people to take up apprenticeships. In China, however, there's also the toy industry, and the fashion accessory industry, so there's a flow of customers, month in and month out. With that come the suppliers and repairmen for the specialized machine tools needed to make the molds, engineering services to design the heating/cooling system so that the plastic will stay liquid to fill the die but harden quickly enough to then be removed. In the US we no longer have the labor force, we don't have the support infrastructure, and we don't have a wide enough customer base to support multiple firms.⁷ Tariffs alone won't do the trick, it has to be a concerted effort that recruits multiple firms to the same place, with the local community college pitching in to recruit and train potential employees.

Trade based on industries with many similar products

There's one other type of specialization, which revolves around the interaction of product differentiation (Toyotas versus Chevys, and pickups versus SUVs versus sedans) and economies of scale, where higher volumes bring lower unit costs. That’s exemplified by BMW. As the manufacturers of premium cars and trucks, any single model achieves only a modest volume in any single country. So BMW and its peers make a specific model in only a single plant. They then export from that plant to the rest of the world. In the case of BMW, its largest factory is in Spartanburg, South Carolina, not in Germany. Spartanburg is also the largest US exporter of cars. As it happens, they make their higher-volume, less-expensive vehicles in the US – we are a large market, so they can sell 40% of what they produce in the US here. They export the remaining 60%. For their higher-end products, no single market is very large, and partly for historical reasons they make those in Germany. The ratio of exports from their German plants is as a result even higher than that of Spartanburg. Again, that’s because it’s cost prohibitive to produce a given model at low volumes in more than one plant. Engine plants with their large castings have even higher economies of scale – in general a car company has only one engine plant for every two to three assembly plants. As a result, when BMW built their large US, it was cheaper to make a modest investment to expand their existing German engine plant, rather than build a small, new one in the US. All the engines in Spartanburg are thus imported.⁸

There are a lot of examples of such "intra-industry" trade. We both import Airbus planes and export Boeings. Automotive steels aren't a commodity, they are a bunch of similar specialized products that require carefully tuned alloys and cold-rolling lines. Move to a different mill, and the subtle changes in iron composition and physical equipment means the steel is different, and the car body panels their customers make don’t come out right. As a result, it's cheaper to make all of one type steel in a single mill and then export to other markets, while importing slightly different steels that are likewise made in a single plant for the global market. Again, car companies tune the stamping dies in a plant to the steel from a specific mill. It will take more than a 25% tariff to justify the cost of building a new mill to move production to the US, plus it takes years to build a new steel mill. Where such specialized plants are located is in part a random legacy of mergers and spinoffs and old plants that got upgraded for new products decades ago. The basic story is that even when the plant-level costs in US and Canada are identical – that is, neither country has a "comparative advantage" – it still makes sense to specialize and trade on that basis. If tariffs make markets smaller, profits disappear – neither steel nor autos are high-margin businesses.

In conclusion

Tariffs won't provide the revenue that the Trump Administration seeks, and won’t lead to reshoring before he leaves office. Building a new factory takes years, and if everyone seeks to do so at the same time, well, there simply aren’t enough industrial engineers or specialized construction contractors to make it happen. Plus we’ve launched a trade war, so global companies that might otherwise contemplate relocating to the US can no longer count on global markets to make a new plant profitable. Instead they’ll face higher input costs and lower competitiveness, and have little assurance that they know what policy will be 6 months from now. They may talk about new investments, but for projects that won’t take concrete form until after November 2028.

Fighting against comparative advantage will in all cases raise costs to US consumers. Higher costs make the US economy less competitive. Tariffs are purportedy MAGA policy, but are actual MAPA policy: Make American Poor Again.

1. In one garment factory I toured in China, most of the workers and supervisors were women, but there were a bunch of men at one group of sewing machines. That image remains vivid, as the factory produced bras.

2. Before the rise of container ships and other logistics innovations, high shipping costs acted as a trade barrier. A small, high-cost plant might survive. In fact, until the 1980s there were still car plants on the east and west coasts. That changed as it became cheaper to ship cars by rail and truck, while with more models on the market, the sales of individual models fell. Car companies no longer needed 2 plants for an individual model. For example, as sales of the Ford Explorer fell, touched off by the Firestone tire rollover scandal, Ford closed its Norfolk VA plant and consolidated production in Kentucky. That had been the last remaining East Coast assembly plant. On the West Coast, the last holdout was GM’s Fremont CA plant. That became the NUMMI joint venture with Toyota, which was at the time buying most components from suppliers in Japan and so wanted a West Coast location. Toyota subsequently bought the plant outright. It closed when Toyota opened its Georgetown KY plant. After sitting vacant for years, Toyota then sold the site to Tesla for a nominal sum – old plants tend to come with an environmental legacy so have low value. Only the building shell was useful, but it had good highway and rail access, electric and water supplies, and a big parking lot.

3. During the covid lockdown in 2020 schools and daycare closed. Both our daughter and her husband were essential workers, but they had two young daughters. Had my wife and I not lived nearby, and been able and willing, my daughter would have had to quit her job, and it was a necessary stepping stone for what proved to be a successful career change. Daycare is a huge need across the US, and when it is available, it changes lives. Where we all live, there are only 6 slots in a licensed daycare for babies, serving a population of 35,000. Pre-K is only slightly better.

4. They were called sweatshops not because the labor was strenuous but because it’s hard to work with cotton when the humidity is low. If the factory walls were sweating, great!

5. This hasn’t made its way into Principles of Economics textbooks. There is already too much material to teach in a two-semester sequence, and faculty don’t want to redo their lecture notes so textbooks are very slow to change. One of the best-selling Principles texts is authored by Paul Krugman and his wife. The international trade chapter makes no mention of his seminal work developing this New Trade Theory that gained him a Nobel Prize in Economics. Specialized “International Economics” texts are only slightly better, because the topic uses different tools than Ricardian comparative advantage, and as a practical matter there isn’t enough time to develop a whole new set of math tools that require calculus rather than graphing.

6. At Harvard it can’t be red, it has to be the very specific Harvard Crimson. Yale Bulldogs wear a particular shade of blue. At Washington and Lee, where I taught, it was a shade of green. Vendors have to see that cotton hoodies, flannel jackets, nylon raincoats, and baseball caps are all the same exact shade.

7. There is a group of tool and die specialists in Windsor Ontario, a few minutes south of Detroit. (Outside of Alaska, it’s the only spot in thousands of miles of border where Canada is to the south, not the north.) Apparently Canada is now an enemy, not an ally, so that can’t serve as a nucleus to develop a robust USMCA industry.

8. BMW did this even though it faced a 10% tariff on the cars it shipped back to Europe. When Audi decided to build a plant here, they instead chose Mexico, which has a free trade agreement with the EU so faces a 0% tariff. That was also true of exports to the US, and to much of Latin America, until Trump decided to unilaterally abrogate the USMCA, a treaty that he in fact negotiated.

To give one more example, MAGA policy includes ending the IRA that supports increasing battery manufacturing capacity, building out charging infrastructure, and subsidizing EV purchases to stimulate volume (via a personal income tax credit of up to $7,500). That might seem to leave GM with a new, empty plant in Orion that could shift to producing full-sized pickup trucks currently assembled in Silao, Mexico. In fact, a truck plant needs more floorspace because the production process is more complicated, while the body-in-white (weld shop) and the assembly line are also specific to an EV, so the machinery can’t be used. Pickup trucks are also too large to fit in the paint ovens. Under current policy, that new plant is now useless and will have to be written off as an impaired asset.