Untitled (ca. 1935–1942) | Farm Security Administration, Office of War Information Photograph Collection, Library of Congress
In 2016 formerly unionized parts of the country, the “Arsenal of Democracy,” now called the Rust Belt (to the dismay of people who live there), switched from voting for the Democratic Party candidate Obama (both in 2008 and 2012) to the Republican Trump. Teresa Ghilarducci and Siavash Radpour (2016) show long-term stagnation and decline of relative economic security in four pivotal Rust Belt states—Michigan, Ohio, Pennsylvania, and Wisconsin. During the Reagan administration in the 1980s, older workers in these four states had much higher wages than their national counterparts. By the time of the Trump/Clinton election in 2016, these states’ older workers’ position relative to the rest of the nation had fallen; they were downwardly mobile in economic security. Wages and employer-provided pensions and health coverage—key barometers of economic security—were lower in these states than the national average. In 1979 older workers in the Rust Belt were making $3,600 more than their counterparts elsewhere. In 2015 they were earning $4,000 less. Between 1979 and 2015, the median real wage for older workers in the four Rust Belt states that flipped to Trump increased only 1 percent compared to growth of 17 percent in the rest of the country.
Scott Abrahams and Frank Levy, in their comprehensive study utilizing “commuting zones” (2023) concluded that the trend toward right-wing political alignment started in the 1980s when deindustrialization began to shift population growth, wealth, and income away from the Midwest. In 1980, 81 of the largest commuting zones accounted for 70 percent of US personal income, but by 2019 income was more concentrated—now 70 percent of it was contained in only 70 commuting zones. Abrahams and Levy (2023, figure 4.4) illustrate the commuting zones where the winning presidential candidate was of a different party in 2016 than in 1980. What is striking about the image is that most of the commuting zones did not change which party they favored. In only 157 out of 722 commuting zones, slightly over 20 percent, the party that won the 2016 presidential election was different from the party that carried that zone in 1980. The map shows coastal commuting zones moving to the Democratic Party—the famous turnover of Republican California to Democratic California over 40 years. The other cluster of the commuting zones that shifted their preference was in the Midwest. But those zones went in the opposite direction, moving from Democratic to Republican.
Further evidence that downward mobility hurts incumbent politicians is that incumbents fare worse electorally in counties with declines in low-wage manufacturing employment and better when employment increases in high-wage sectors. The research does not establish a causal relation, but it is not surprising that immiserated people—relative to their reference group—would vote for change (Jensen, Quinn, and Weymouth 2017). Evidence from Germany suggests economic insecurity leads to right-wing voting (Bossert et al. 2019). The upward spiral of populism in Europe has been attributed to economic insecurity (Guiso et al. 2024). Leo Kahane (2020) finds that US counties with increased poverty rates (and suburban counties) were more likely to support the Republican candidate in 2016 compared to 2012. Income inequality, the share of the Black male and female populations, Hispanic male population, and percent of the population with a college degree all had a larger effect on the explanation of the Democratic vote in 2016 than in 2012.
In sum, the literature on voting suggests that relative material well-being and changes in happiness are more important for explaining changes in voting behavior than absolute levels of well-being and happiness. Declining relative economic status creates more unhappiness, which in turn leads to behavior protesting an incumbent regime or party. We acknowledge that other factors are linked to a growing attraction to populism, like a falling confidence in core democratic institutions, which, too, may be related to a drop in relative economic security. We conclude that research on voting should focus on place and community and the status of that place relative to other places. There is no countywide measure of subjective well-being. Our hypothesis is that a certain kind of economic anxiety—a community’s downward mobility—explains voting patterns of a county better than absolute levels of well-being or poverty.
