Free Trade Bulletin No. 3

The U.S. Trade Deficit and Jobs: The Real Story

By Daniel Griswold
February 3, 2003

On February 20, 2003, the U.S. Department of Commerce reported that the U.S. trade deficit reached a new record in 2002. For the calendar year, imports of goods into the United States exceeded exports by $484.4 billion. When that figure is combined with an overall surplus in services of $49.1 billion, the 2002 deficit in goods and services was $435.2 billion, the largest in U.S. history.[1] If the past is any guide, the record deficit will be misused by opponents of free trade to claim that it depresses production and destroys jobs in the U.S. economy. The argument of the trade critics is simple but misleading: If exports create jobs, then imports must destroy jobs. Thus, a trade deficit by its very nature causes a net loss of jobs in the U.S. economy, and the bigger the deficit, the more jobs lost. The leading proponent of this sort of analysis is the Economic Policy Institute, a pro-union, left-of-center nonprofit organization in Washington that routinely publishes studies claiming to show specific job losses, including state-by-state totals, as a result of trade deficits. Those numbers are then repeated by labor union leaders and trade opponents in Congress to warn against further trade liberalization. As the author of one EPI study explains: “When the United States exports 1,000 cars to Germany or Mexico, plants in this country employ U.S. workers in their production. If, however, the U.S. imports 1,000 cars from Germany or Mexico rather than building them domestically, then a similar number of U.S. workers who would have otherwise been employed in the auto industry will have to find other work. Ignoring imports and counting only exports is like balancing a checkbook by counting only deposits but not withdrawals.” [2] To determine the number of jobs or potential jobs “eliminated” by the trade deficit, the EPI model compares actual U.S. employment to what it would supposedly be if the U.S. trade deficit were zero and the economy’s overall growth rate unchanged. Fewer imported cars, steel slabs, shoes, toys, shirts, and other goods are then translated into more domestic production of those items and hence more jobs if exports equaled imports. In other words, every widget not imported translates into a widget produced at home and more widget workers employed. Within this model, the rising imports and trade deficits of recent years can only be bad news for output and employment. As EPI concludes, “The toll on U.S. employment has been heavy: from 1994 to 2000, growing trade deficits eliminated a net total of 3.0 million actual and potential jobs from the U.S. economy.”[3] The attempt to blame trade deficits for a loss of jobs founders in theory and in practice. First, the model ignores the role of international investment flows. The flip side of America’s trade deficit is the net inflow of foreign investment. The extra $435 billion that Americans spent on imports over and above exports last year was not stuffed into mattresses overseas. Those dollars quickly returned to the United States to buy U.S. assets, such as stocks, bank deposits, commercial and Treasury bonds, or as direct investment in factories and real estate. A principal reason why the United States runs a trade deficit with the rest of the world year after year is that foreign savers continue to find the U.S. economy an attractive place to invest. The EPI model ignores the growth and jobs created by the offsetting inflow of net foreign investment into the U.S. economy that the trade deficit accommodates. That net surplus of investment capital buys new machinery, expands productive capacity, funds new research and development, and keeps interest rates lower than they would otherwise be. EPI counts the jobs supposedly lost when we import cars but ignores the jobs created when BMW or Toyota builds an automobile factory in the United States that employs thousands of Americans in good-paying jobs. So it is the critics of trade who are guilty of counting the withdrawals but not the deposits in our national balance of payments account. Second, the central assumption of the EPI model—that rising imports directly displace domestic output—collides headlong with empirical reality. In fact, imports and domestic output typically rise together in response to rising domestic demand. During much of the 1990s, when imports and trade deficits were both rising rapidly, so too was domestic employment and manufacturing output. Between 1994 and 2000, when deficits supposedly claimed a “heavy toll” on U.S. employment, civilian employment in the U.S. economy rose by a net 12 million[4] and the unemployment rate fell from 6.1 percent to 4.0 percent.[5] During that same period, U.S. manufacturing output rose by 40 percent[6] even though the volume of imported manufactured goods doubled.[7] Manufacturing took a nosedive in 2001-2002, but rising imports were not the culprit. While manufacturing output was falling 4.1 percent in 2001 from the year before,[8] real imports of manufactured goods were falling 5.4 percent after four straight years of double-digit increases.[9] The same domestic recession that put the kibosh on domestic manufacturing output also curbed demand for imports. In fact, imports and output have been tightly and positively linked in recent years. The scatter-plot chart reveals the remarkable correlation between the growth of manufacturing imports and manufacturing output each year since the late 1980s.

U.S. Manufacturing Output and Imports, 1989-2002 In the chart, output growth is measured by the change in the average level of output for the year compared to the average level of output in the previous year. Import growth is measured by the total volume of manufactured goods (capital goods, industrial supplies and materials, automotive vehicles, engines, and parts, and durable and non-durable consumer goods) imported in that year compared to the volume imported the previous year. The chart illustrates that the more we import, the more we make ourselves; the more we make, the more we import. Essentially, years fall into one of two categories: high-import, high-output years, or low-import, low-output years. During the late 1980s and early 1990s, as the U.S. economy slowed, the growth of both manufacturing imports and output slowed. As the expansion of the 1990s gained steam, growth of imports and output accelerated. Then the recession of 2001 and the slow recovery in 2002 saw imports and output slide down the scale together. If the trade critics were right, the recent plunge in import growth should have stimulated an increase in domestic output as U.S. factories sought to fill the gap left by the missing imports. According to the EPI model, in other words, the relation should be negative and the trend line should slope downward and not upward. Once again, reality intrudes on the protectionist story. There is no basis, in theory or experience, for the persistent allegation that trade deficits, and more specifically imports, mean fewer jobs in the U.S. economy. The reality is more nearly the opposite. As a reflection of continued domestic demand and the desire of foreign investors to acquire U.S. assets, large trade deficits are typically associated with more output and more jobs. In America today, trade and prosperity are a package deal. The more we trade, the more we prosper, and the more we prosper, the more we trade. By seeking to curb imports of manufactured goods, opponents of trade will only undermine the ability of the U.S. economy to expand output and create jobs.


[1] U.S. Bureau of the Census, “U.S. International Trade in Goods and Services: December 2002,” Report Text, February 20, 2003, p. 3, www.census.gov/indicator/www/ustrade.html.

[2] Robert E. Scott, “Fast Track to Lost Jobs: Trade Deficits and Manufacturing Decline are the Legacies of NAFTA and the WTO,” Briefing Paper, Economic Policy Institute, October 2001, p. 2.[3] Ibid., p. 1.[4] Total civilian employment rose from 123.1 million in 1994 to 135.2 million in 2000. Council of Economic Advisers (CEA), Economic Report of the President 2003 (Washington: Government Printing Office, February 2003), Table B-36, p. 320.[5] Ibid., Table B-42, p. 326.[6] The Federal Reserve Board’s index of U.S. manufacturing output rose from 83.7 in 1994 to 117.4 in 2000. Ibid., Table B-51, p. 336.[7] Real imports of manufactured goods increased from $541.6 billion in 1994 to $1,105.7 billion in 2000. U.S. Department of Commerce, Bureau of Economic Analysis (BEA), “National Income and Product Account Tables,” Table 4.4. Real Exports and Imports of Goods and Services by Type of Products [Billions of chained (1996) dollars], Revised January 30, 2003, www.bea.doc.gov/bea/dn/nipaweb/SS_Data/Section4All_xls.xls.[8] CEA, Table B-51, p. 336.[9] BEA, Table 4.4.

Read the Full Free Trade Bulletin

Daniel Griswold is the director of the Center for Trade Policy Studies at the Cato Institute.