Correcting the Record on NAFTA



In a recent blog, Kevin P. Gallagher of Boston University takes the Peterson Institute to task for “overselling” the North American Free Trade Agreement (NAFTA) to the American public in 1993. To make this criticism, Gallagher lifts a phrase from our assessment published prior to NAFTA ratification: “US exports to Mexico will continue to outstrip Mexican exports to the United States.”1

In NAFTA: An Assessment (1993, p. 14), we wrote:2

Our job projections reflect a judgment that, with NAFTA, US exports to Mexico will continue to outstrip Mexican exports to the United States, leading to a trade surplus with Mexico of about $7 billion to $9 billion by 1995. How long can that scenario last? The answer fundamentally depends on investor confidence in Mexico, Mexican growth, and the ratio between Mexican external debt and Mexican GDP.

Gallagher conveniently ignored the last two sentences in this forecast. As we now know, investor confidence in Mexico collapsed at the end of 1994, owing to excessive dollar-denominated debt, resulting in a drastic devaluation of the peso. As a consequence, Mexican exports to the United States surged, leading to a bilateral merchandise trade deficit of $16 billion in 1995.

But the larger point is quite different. Unlike Gallagher and other critics, we never regarded the bilateral trade surplus or deficit as a decisive litmus test for NAFTA’s success. For us, the key test was two-way trade growth. In other words: Would NAFTA deliver faster trade growth with Mexico than might have otherwise occurred?

We projected that NAFTA would deliver a long-term increase in two-way US merchandise trade with Mexico of $30 billion, in each direction (measured in 1990 dollars), above the normal growth in world trade after 1990. Measured in 2005 dollars, this projection suggested that the long-term NAFTA payoff would have been $43 billion in additional trade in each direction within a decade. However, by 2005, US merchandise exports to Mexico were $53 billion more than what might have been expected from the growth of US exports to the world, and US merchandise imports were $55 billion more.3 The figure below shows the course of US-Mexico trade since 1990. Calculations explained elsewhere indicate that the NAFTA contribution to US two-way trade with Mexico added $26 billion annually to US GDP as of 2005 and more in successive years.4

​​​​​​Contrary to Gallagher and other critics, both imports and exports benefit the American public. The multiple payoff channels are summarized in chapter 2 of The United States and the World Economy (Peterson Institute for International Economics, 2005). As for imports, critics focus almost exclusively on their job displacement effects and ignore the benefits of lower prices, higher productivity, and greater variety. Far more jobs are displaced by technology than trade, and US imports from Mexico are around 13 percent of US imports from the world. Nevertheless, based on the records of the NAFTA Trade Adjustment Assistance (TAA) program, some 414,000 workers were displaced between 1994 and 2002.5 NAFTA displacement represented less than 2 percent of total job displacements during the same period, about 31 million workers—over 3 million a year. Despite predictions of NAFTA critics that the trade agreement would destroy jobs, the US economy grew significantly in the first decade of NAFTA with annual growth of 3.3 percent.6 That’s because trade agreements have little net effect on the level of national employment but do contribute to the overall “churn” in the job market. Existing government programs are totally inadequate to the hardships faced by displaced workers. The answer is not to denounce NAFTA and other trade agreements but to create the kind of adjustment programs that exist in other advanced countries.

Another criticism of NAFTA is wage depression. A study of national metropolitan labor markets by Autor, Dorn, and Hanson (2013) was unable to detect an impact on wages from US trade with Mexico and Central America.7 However, a subsequent study of local labor markets by Hakobyan and McLaren (2016) did find that about 1.3 million American workers experienced slower wage growth of 5 percent or more during the 1990s because of import competition from Mexico.8 These NAFTA-impacted workers represented about 1 percent of US workers in 2000. Their losses are real, but they are substantially outweighed by broad gains to the American economy.

Gallagher’s overarching complaint is that corporations essentially wrote the original NAFTA and silenced other constituents. His biggest fear is that corporations will dominate the NAFTA renegotiation. Gallagher’s misreading of history ignores President Clinton’s labor and environmental side agreements, and his crabbed forecast ignores President Trump’s populist leanings. The necessity of congressional approval will ensure that the new NAFTA, like the original NAFTA, reflects the preferences of the American public.

Authors' note: We thank research analyst Euijin Jung for valuable assistance.


1. Gallagher also suggests that we claimed that Mexico would finally “export goods, not people.” While that is not a phrase we can find in our writings, it echoes Mexican President Carlos Salinas who famously framed NAFTA as a “choice between getting Mexican tomatoes or tomato pickers.”

2. Gary Clyde Hufbauer and Jeffrey J. Schott, NAFTA: An Assessment, Peterson Institute for International Economics, Washington, February 1993.

3. Measured in 2005 dollars, US merchandise exports to Mexico were $120 billion in 2005, and US merchandise imports from Mexico were $170 billion. If US two-way trade with Mexico had grown between 1990 and 2005 only as fast as US trade with the world, US merchandise exports would have been just $67 billion, and US merchandise imports would have been only $115 billion.

4. For a detailed analysis of the successes and shortcomings of NAFTA, see Gary Clyde Hufbauer, Cathleen Cimino-Isaacs, and Tyler Moran, NAFTA at 20: Misleading Charges and Positive Achievements, Policy Brief 14-13, Peterson Institute for International Economics, May 2014.

5. The NAFTA TAA program was merged into the general TAA program after 2002.

6. Gary Clyde Hufbauer and Jeffrey J. Schott, NAFTA Revisited: Achievements and Challenges, Peterson Institute for International Economics, Washington, October 2005.

7. David H. Autor, David Dorn, and Gordon H. Hanson, The China Syndrome: Local Labor Market Effects of Import Competition in the United States, NBER Working Paper No. 18054 (May), Cambridge, MA: National Bureau of Economic Research, 2013. This paper, and subsequent work by the same authors, found a measurable negative effect of Chinese imports on US wages.

8. Shushanik Hakobyan and John McLaren, “Looking for Local Labor Market Effects of NAFTA,” Review of Economics and Statistics, vol. 98, no. 4, October 2016. The figure of 1.3 million comes from a calculation supplied by professors McLaren and Hakobyan to the authors from their analysis.