The sensational title of a recent Working Paper published by the National Bureau of Economic Research (NBER) suggests that the North American Free Trade Agreement (NAFTA) did kill workers: Trading Goods for Lives: NAFTA’s Mortality Impacts and Implications.
The authors compared mortality in local labor markets exposed to competition from Mexico after NAFTA took effect in 1994 with mortality in other labor markets. They found a 0.68 percent annual increase in age-adjusted mortality over the following 15 years and attributed the entire increase to NAFTA’s impact on manufacturing jobs.
The New York Times was so convinced of NAFTA’s killing power that it ran an even more sensational title in March: “A Lot of Life Years Lost’: How NAFTA Shortened American Life Spans.”
But the Working Paper’s authors—Amy Finkelstein and Steven X. Shi of MIT, and Matthew J. Notowidigdo of the University of Chicago—erred in significant ways that fueled the exaggerated headlines.
The authors’ econometric specification did not allow for the possibility that NAFTA may have reduced nationwide mortality even if it did not reduce mortality among those who lost manufacturing jobs as much as the rest of the population.
But more important, statistical association is not the same as policy causation. The failure of US policy was not the creation of NAFTA. The failure was, and remains, the absence of a meaningful safety net for displaced workers.[1]
While presented as a provocative critique of NAFTA, the NBER Working Paper’s statistical findings are hardly new. In 2006, a working paper published by the Federal Reserve Bank of Chicago used quarterly earnings data to show that displaced manufacturing workers experienced a 15-20 percent increase in death rates over the following 20 years.
And the Chicago Fed paper was just one contribution to a large body of statistical literature on the association between job loss and mortality. Meta-analysis of 42 studies on mortality risk published in 2011 by the journal Social Science & Medicine found that mortality following job loss increased by 63 percent after taking into account multiple characteristics of individual workers.
What the NBER authors did in their title was to cherry-pick one cause of job loss, namely imports from Mexico, and suggest that NAFTA was uniquely responsible for “deaths of despair” (to cite the book by Anne Case and Angus Deaton).[2] The body of their paper, however, reports that manufacturing job loss, for whatever cause—trade, recession, technology—was statistically associated with higher mortality. There is nothing particularly unique about NAFTA, or imports from Mexico, or imports in general.
Between January 1994 and December 2008 (the period examined by the NBER Working Paper) US manufacturing jobs declined from 16.9 million to 12.8 million. The impact of NAFTA and imports from Mexico accounted for a small part of that loss of 4.1 million manufacturing jobs. To the extent imports are a source of job loss, between 1994 and 2008 US imports of manufactures from Mexico were only 10.2 percent of total US imports of manufactures.[3] Moreover, as Robert Z. Lawrence has documented, technology and shifting consumer tastes are more important sources of declining manufacturing jobs than imports. Meanwhile, the opioid epidemic, greatly expanding the supply of death-dealing drugs, erupted in the late 1990s, around the same time as larger manufactured imports from Mexico. Opioid consumption triggered by the loss of manufacturing jobs and abrupt earnings losses is not separately explored in the NBER paper.
In any event, policy causation deserves the focus of attention, not statistical association. Based on research by Lawrence, manufacturing jobs were destined to disappear, but the ultimate cause of increased mortality was the failed US policy response. Yet the NBER paper has little to say about policy causation. Other research is more informative.
The cited Chicago Fed paper found that the sharp decrease in average earnings following job loss was a major factor explaining higher mortality. A very good paper by Kerwin Kofi Charles et al. in 2018 supports the idea that it is the loss of manufacturing jobs, not imports, that is responsible for earnings loss. Focusing their attention of the “shock” of imports from China, they find that “the local labor market effects of local manufacturing employment due to Chinese import competition are very similar to the local labor market effects of manufacturing declines due to other forces.”
Turning to the policy consequences of earnings loss, research comparing US and German mortality following job loss concluded that the much better German social safety net largely explains the much lower German mortality rate. Other studies report similar findings.
Considering the huge benefits from freer trade, it makes no sense to answer “deaths of despair” by abolishing NAFTA, its reincarnation in 2020 as the United States-Mexico-Canada Agreement, or other free trade agreements. Nor does it make sense to halt technological progress by guaranteeing lifetime jobs to manufacturing workers.
What does make sense is to repair the shoddy US social safety net for displaced workers. Temporary compensation for lost earnings and health care should be provided to workers who lose jobs through no fault of their own. Proposals have been floated for wage insurance, guaranteed minimum income, and universal basic income, but none have been enacted. The sole significant improvement in the US safety net in this century, the 2010 Affordable Care Act, is now threatened.
The real US policy failure is the absence of European-style programs for displaced workers. Manufacturing job loss is yesterday’s story. It remains to be seen whether this US policy failure will be addressed before the artificial intelligence–tsunami displaces a large number of service sector workers.
Notes
1. See C. Fred Bergsten (1975), The Dilemmas of the Dollar: The Economics and Politics of United States International Monetary Policy. George Shultz, then secretary of labor, decided in 1969 that it would be too costly to extent adjustment assistance to all displaced workers, and the trade adjustment assistance program (TAA) was accordingly limited to trade-displaced workers. But even for trade-displaced workers, TAA proved inadequate.
2. Anne Case and Angus Deaton, Deaths of Despair and the Future of Capitalism, Princeton University Press, 2020.
3. Manufactured imports include SITC 5 (Chemicals and related products, n.e.s.), SITC 6 (Manufactured goods classified chiefly by material), SITC 7 (Machinery and transport equipment), SITC 8 (Miscellaneous manufactured articles). The data source is US International Trade Commission DataWeb.
Data Disclosure
This publication does not include a replication package.
