Reassessing the Backlash of Manufacturing Job Loss in Presidential Elections

Many see deindustrialization as having generated a politically active group of embittered and reactionary voters motivated by their declining social position as ‘globalization losers’ as a result of their income, education, class or occupation. As the backlash narrative became more widespread amongst scholars and political pundits, a conventional wisdom emerged that the loss of manufacturing jobs in the American economy has spurred support for the Republican party in places which have been 'left behind' in the adaptation to a service-based economy. The aim of this study critically evaluate whether manufacturing job loss increased changed electoral politics, and in particular has it increased support for the Republican party. To do so, we pool electoral results with local administrative data from 1980-2016. Our study shows finds no evidence that manufacturing job loss has impacted the relative popularity of the Republican or Democratic party in US presidential elections from 1980 to 2016. In a cross-sectional analysis at the county level we also find no significant relationship between the decline of manufacturing employment and the vote for Trump This null result is robust to various model specifications and different measurements of deindustrialization. Our findings show the importance of the long-run to study the presumed consequences of deindustrialization. While politicized plant closures, manufacturing employment decline, and the nostalgia for an industrial past are a convenient explanation, we argue that it does not have an empirical basis and question that the generalized conclusion that this could be fueling populist support in other contexts as well.