The assumed connection between rising imports and falling production sounds plausible. If Americans buy more shirts, shoes, and TV sets from producers abroad, it would seem obvious that American companies would then produce fewer shirts, shoes, and TVs to meet domestic demand. Mass layoffs and closed factory gates would soon follow, or so we are told by certain politicians and cable TV commentators.
That protectionist claim can be tested by comparing the volume of manufacturing imports to manufacturing output from year to year. If the critics of trade are correct in their assumptions, a rise in the growth of manufacturing imports should lead quite directly to a decline in the growth of manufacturing output. By the same reasoning, a decline in imports should stimulate domestic output, as consumers substitute domestic‐made goods for foreign‐made goods. But an analysis of manufacturing imports and output since 1989 plainly refutes this pillar of protectionist thinking.
Figure 1 shows the change from the previous year in real (inflation‐adjusted) manufacturing imports and manufacturing output in the United States for each year since 1989. Manufacturing imports are defined as industrial supplies and materials, capital goods, automotive vehicles and parts, and consumer goods. Manufacturing output is measured by the annual average of the Federal Reserve Board’s monthly index of manufacturing output. The percentage change in real manufacturing imports from the previous year is plotted on the horizontal axis, and the percentage change in manufacturing output from the previous year is plotted on the vertical axis.
If the critics of trade were correct, the trend‐line would be sloping down–that is, the more rapidly imports grow in a particular year, the more depressed we would expect manufacturing growth to be in that same year. A dreaded “flood of imports” should mean slower growth or outright contraction of manufacturing output, while a slowdown of import growth should bring relief to domestic producers and thus spur faster domestic output. In reality, the opposite proves to be true.
Since 1989, accelerated growth in real manufacturing imports has been strongly tied to accelerated growth in domestic manufacturing output. Indeed, the correlation is almost uncanny, with no year straying far from the trend‐line. Years of rapid growth in manufacturing imports are also years of rapid growth in manufacturing output, and years of slower growth in imports are years of sluggish growth, or even declines, in domestic output.
Specifically, the years 1989–91 and 2001–2003 were periods of low growth in both imports and output. In contrast, the years 1992–2000 saw robust manufacturing output growth coupled with robust manufacturing import growth. Consider two recent years at opposite ends of the line: In 2001, manufacturing output fell in the United States by 4.0 percent, but the fall in manufacturing imports was an equally sharp 4.7 percent. In 2004, manufacturing output rebounded with strong growth of 4.8 percent, accompanied by strong import growth of 12.5 percent.
One explanation for the positive link between imports and output is that American producers themselves are major importers. In fact, about half of the goods imported into the United States each year are not for consumers directly, but for U.S. businesses. American firms import capital machinery, industrial supplies and materials, and petroleum and other raw materials that allow them to produce final products for sale to consumers. As American‐based companies ramp up production to meet rising consumer demand, they must import more capital goods, intermediate inputs, and raw materials to keep the assembly lines humming.
Another reason why imports and output rise together is that, in a flexible, market‐driven domestic economy, resources can quickly shift from one sector to another. When imports displace domestic production in one sector, it frees resources to expand production in other sectors of the economy where Americans can produce goods and services more efficiently. Thus, imports do not lead to less output and fewer jobs, but to a shift in resources, production, and employment to sectors where Americans are more productive.
Indeed, for U.S. manufacturers, imports and output are a package deal. Rising domestic demand and output spur imports, while imports promote gains in productivity and output.
The net loss of jobs in manufacturing cannot be blamed on rising imports, because imports cannot be blamed for falling manufacturing output. In fact, the largely positive story of U.S. manufacturing during the past decade has been one of rising output produced by fewer workers. In other words, the output per manufacturing worker has been rising so rapidly that it has outstripped the rise in production, allowing U.S. factories to produce more output with fewer workers.
Although manufacturing imports do not depress total output, they do accelerate the shift of output into sectors where American workers are even more productive. Between 1989 and 2004, the volume of manufacturing output in the United States grew an impressive 51 percent, but behind that overall growth was a significant shift in what Americans manufacture.
Since 1989, the volume of output has grown dramatically in such high‐tech sectors as semiconductors, printed circuit boards, and other electronic components; computers; and video, audio, and other communications and information‐processing equipment. Output has also grown more rapidly than average in recreational vehicles, campers, automobiles, light trucks, and automotive parts; medical equipment; pharmaceuticals and medicine; and chemical products.
During that same period, other sectors of U.S. manufacturing have seen dramatic declines in output. Most prominent among the shrinking industries are apparel and leather goods, textile, fabric and yarn mill products, newspaper publishing and paper mills, logging, primary smelting, nonferrous metals, pig iron, and coke. Also contracting during that period, especially in the late 1980s and early 1990s, were the aerospace, aircraft and parts, and defense and space equipment sectors–not primarily because of rising import competition, but because of the end of the Cold War.
Import competition has undoubtedly accelerated the decline of a number of those shrinking, sunset industries. The most vulnerable to import competition are labor‐intensive industries where the lower wages paid in developing countries allow their producers to compete most effectively. But growing trade also allows Americans to shift production into more capital‐intensive sectors where more skilled and educated Americans enjoy a comparative advantage. As a result, trade is not about more or fewer jobs and output, but about better jobs and more output in those sectors where we enjoy the most advantages as a nation.
The evidence makes clear that protectionists are selling a vision that does not reflect reality. They believe that if imports are reduced, through tariffs or currency adjustments or other policy tools, Americans will buy more domestically produced goods instead, and create more and better paying jobs at home. But the reality of the American economy is closer to the opposite. The protectionist dream is really a nightmare for U.S. manufacturers. Slower growth of imports typically means slower growth in domestic output.
The evidence from the past 15 years indicates that raising trade barriers to supposedly protect U.S. manufacturing would be a mistake. Raising barriers would not stimulate overall production. It could save jobs in certain industries, at least temporarily, but it would do so at the expense of output in other, generally more competitive industries. Protectionism would only slow the ongoing, beneficial transition within U.S. manufacturing to those sectors where Americans are better able to compete in world markets.
 U.S. Commerce Department, “National Income and Product Accounts,” Bureau of Economic Analysis, Table 4.2.6., Real Exports and Imports of Goods and Services by Type of Product, www.bea.doc.gov/bea/dn/nipaweb/index.asp.
 U.S. Federal Reserve Board, “Industrial Production and Capacity Utilization: Historical Data,” Industrial Production, Seasonally Adjusted, Tables 1 and 2, www.federalreserve.gov/releases/g17/ipdisk/ip.sa. In statistical terms, the correlation is 0.94, with 1.00 being a perfect correlation. See U.S. Federal Reserve Board, “Industrial Production and Capacity Utilization: Historical Data.”