Matthew C. Klein and Michael Pettis—

The drive south across the Detroit River from the U.S. city of the same name to the Canadian city of Windsor only takes about twenty minutes. For nearly a century, the big three American automakers have exploited this proximity by operating plants in both Michigan and the southern bit of Ontario. This Great Lakes motor vehicle manufacturing complex might be the first recognizably modern global value chain. Components, material inputs, and finished cars and trucks are constantly moving across the border. Trade between Canada and Michigan is more valuable than trade between Canada and China. Trade in motor vehicles and parts between the United States and Canada is worth more than a fifth of the total trade between the two countries.

Thanks to the collapsing cost of transportation, especially by sea, cross-border manufacturing networks are now far more common and far more spread out than in the years when Chrysler, Ford, and General Motors first opened factories and offices in Canada. Most of the world’s manufacturing takes place in one of three cross-border manufacturing networks centered on the United States, Germany, and China (previously Japan until about 2007). Trade of intermediate inputs within these networks explains more than half of all international trade, while cross-border trade of finished goods and services accounts for only one-third. (Energy and metals commodities make up the rest.) This is a far cry from the world of Smith and Ricardo or even the world of the 1960s. It is a consequence of containerization, liberalization, and the end of the Cold War.

American exports of goods to Canada and Mexico are worth about as much as U.S. exports of goods to the European Union, China, Japan, and Korea combined. Much of the value of U.S. exports to its neighbors, however, comes from elsewhere. A seatbelt for an American-made car or light truck, for example, might have its fibers manufactured in Mexico, woven and dyed in Canada to take advantage of the abundance of water, sent back to Mexico to be sewn up, and then installed somewhere at a plant in the United States. At the same time, nearly half of the content of motor vehicles and parts imported from Mexico was originally made in the United States.

Trade in goods and services among the twenty-eight members of the European Union (before Brexit) was worth about 50 percent more than trade between those same countries and the entire rest of the world. The German-led motor vehicle supply chain stretches east into Czechia, Hungary, Poland, Romania, and Slovakia, and southwest into Portugal and Spain. Nearly half of the value of exports sold by Germany’s eastern neighbors comes from foreign components. More than half the motor vehicles produced by German car companies are made outside Germany, and about a third of the value of Germany’s own automotive exports comes from its neighbors.

Perhaps the most iconic transnational supply chain is the one developed to assemble electronics in China. In 2007, China ostensibly exported about $290 billion worth of “computer, electronic, and optical” products, but roughly $120 billion of the value of those exports (around 40 percent) came from elsewhere, particularly Korea, Japan, and Taiwan. Even though Chinese producers have since become far less reliant on imported components, global value chains still have an impact on the bilateral trade data. Even now, about a third of the value of China’s imports from Korea and Taiwan originated elsewhere, reflecting those countries’ positions in the middle of international supply chains. Taiwanese academics calculate that standard figures overstate the value of their country’s trade relationship with China by a factor of three.

The increasing importance of these global value chains means that conventional bilateral trade data no longer do a good job of measuring the actual value created by workers and machines in each country. Gadgets assembled in China (or, nowadays, Vietnam) and shipped to North America or Europe are filled with imported components, including components made in the United States, just as German cars are built with Eastern European parts and American trucks are filled with Mexican content. Yet the statistics produced by customs offices attribute all of the value of the imported inputs to whichever country happens to ship the finished product. Economists have recently begun producing alternative trade statistics that account for these transnational manufacturing networks. For the United States, imports are overstated by about 16 percent while exports are overstated by about 20 percent. Chinese imports and exports are both overstated by about 30 percent.

The growth of transoceanic shipping among the three main manufacturing networks adds to the confusion. Although governments are reasonably good at tracking where goods came from and where they are initially shipped, it is much tougher for the customs offices to follow exports to their final destinations. American exports to the European continent often land in the major ports of Antwerp or Rotterdam before being moved to the major markets of France, Germany, and Italy. Similarly, many U.S. exports arrive in Hong Kong and Singapore before moving elsewhere in East Asia.

One strange result is that the United States consistently reports massive exports to Belgium, the Netherlands, Hong Kong, and Singapore. Official U.S. data imply that American businesses supposedly exported about $151 billion worth of goods to those four small countries in 2018. That would be more than American goods exports to China ($121 billion), or about as much as total U.S. goods exports to France, Germany, and the United Kingdom combined ($161 billion). Put another way, U.S. exports of goods to Belgium and the Netherlands ($80 billion, combined population of 29 million people) were ostensibly worth as much as exports sent to Germany and Italy together ($81 billion, combined population of 140 million). Meanwhile, the value of American goods shipped to Hong Kong and Singapore ($71 billion, combined population 13 million) was just below the value of U.S. exports of goods to Japan ($76 billion, population of 127 million). These numbers are artifacts of flawed international trade reporting rather than serious economic indicators.

Combined, transnational manufacturing networks and oceanic ports mean that official figures on exports and imports are a poor guide to the countries that are capturing the profits and employment income from international trade.

From Trade Wars Are Class Wars by Matthew C. Klein and Michael Pettis. Published by Yale University Press in 2021. Reproduced with permission.

Matthew C. Klein is the economics commentator at Barron’s. Michael Pettis is professor of finance at Peking University’s Guanghua School of Management and a senior fellow at the Carnegie Endowment for International Peace.

Further Reading: