The Politics of Manufacturing Decline

County colors represent:

• Red	Greater than median of vote shift, and greater than median of job loss.
• Pink	Greater than median of vote shift, and less than median of job loss.
• Light Blue	Less than median of vote shift, and greater than median of job loss.
• Dark Blue	Less than median of vote shift, and less than median of job loss.

The Issue:

Workers in declining manufacturing regions of the United States have been among the strongest supporters of the populist upsurge that includes a rejection of globalization. Many studies link the upsurge in hostility to trade to changes in economic conditions, especially the loss of manufacturing jobs (Jensen, Quinn, and Weymouth 2017, Freund and Sidhu 2017, Autor et al 2017, Che et al 2016). As a presidential candidate, Donald Trump blamed globalization for the decline of American manufacturing, arguing “Globalization has left millions of our workers with nothing but poverty and heartache,” calling NAFTA “the worst trade deal in history,” and charging that “China’s entrance into the World Trade Organization has enabled the greatest job theft in the history of our country.”  Indeed, many observers ascribe the victory of Donald Trump in the 2016 presidential election to his appeal to disaffected white workers in the country’s manufacturing areas.

Is there, in fact, an association between the decline of traditional manufacturing and the rise of the populism embodied by Donald Trump?

The Facts:

• Manufacturing employment has declined in the United States. Workers in manufacturing represented 26 percent of the labor force at the beginning of 1970, and this has decreased to 8.5 percent at the beginning of 2019. This trend is common across industrial countries, including Germany, Great Britain and Japan. Technological change was an important reason for this trend, but imports from low-wage developing countries and the movement of factories to lower-wage countries also contributed to the loss of many manufacturing jobs in advanced countries. (Acemoglu and Restrepo 2017, Pierce and Schott 2016, Autor, Dorn, and Hanson 2013)
• In the United States, the loss of manufacturing jobs has been concentrated in the traditional Industrial Belt, from upstate New York and western Pennsylvania through the Great Lakes states to the upper Midwest. Many communities in this region have experienced decades of industrial decline (Fort, Pierce, and Schott 2018). The loss of well-paying manufacturing jobs has sent many of the small cities and towns in this area into a downward spiral. As factories closed, wages and income declined and brought the local tax base down with them. Local public services – schools, infrastructure, and law enforcement – suffered. [Feler and Senses 2017, Che, Xu, and Zhang 2018). Soaring housing prices in the prosperous cities made it difficult for struggling families to leave depressed areas and move to the booming cities (Ganong and Shoag 2017). As those who could migrated, the region was left with an aging, less educated workforce. 
• There is substantial evidence for the electoral importance of disaffected white workers in the 2016 election, especially in the Industrial Belt, who saw in the Trump campaign a response to their problems. (Baccini and Weymouth 2018, Ballard-Rosa, Jensen, and Scheve 2018). The map presents the result of our county-level statistical analysis of voting and manufacturing job loss that helps illustrate the importance of manufacturing decline to the election of Donald Trump. The counties that are colored red represent those in which there was a bigger than median-value shift in both votes for Donald Trump in 2016 from those in 2012 for Mitt Romney , a mainstream Republican, and in the loss of manufacturing jobs (this “Trump shift” takes several other factors into account, including average age, average education, and the racial composition of these counties). The counties colored dark blue had the opposite profile, smaller-than-median shifts in voters towards Trump and smaller than median losses in manufacturing jobs. The counties colored pink had larger than median shifts in voting for Trump and smaller than median losses in manufacturing, while the counties colored light blue had smaller than median shifts in voting for Trump and larger than median shifts in manufacturing job losses. (The counties colored grey were those for which we did not have sufficient data to conduct this analysis.) It is readily apparent that most of the relevant political change was in the country’s manufacturing belt. In the rest of the country, either the decline in manufacturing and the Trump shift were not particularly substantial, or they were not related to each other. 
• The loss of manufacturing jobs is one factor that contributed to the victories of Donald Trump in the Republican primaries and the 2016 general election but, of course, there are others as well. The Great Financial Crisis of 2008-2009, the longest and deepest economic downturn since the 1930s, hit the American middle class especially hard. There were important ethnic and racial components to Donald Trump’s appeal, especially with respect to immigration and, more generally, with the changing ethnic and racial makeup of American society. The election also reflected the growing gap between the social and political attitudes of Americans in rural areas and the country’s small cities and towns, on the one hand, and large urban areas, on the other. Finally, political differences between more and less educated Americans have increased: college-educated whites voted heavily for Hillary Clinton, while whites without a college education voted heavily for Donald Trump.