American Deindustrialization, American Chaos
American manufacturing was a major political nerve in our recent Presidential election, in which the winner earned unexpected votes in many regions that have seen manufacturing jobs leave for foreign shores. President Trump did this by demonizing the political status quo, blaming them for the financial crisis, the ensuing austerity, and neoliberal trade policies that caused jobs to flee. These criticisms were largely incoherent, yet the nerve was pressed nonetheless. But while trade policy has played a significant role, nearly of our country’s anxiety about manufacturing jobs can be traced back to the longer phenomenon of deindustrialization.
Deindustrialization is a well-documented trend in the developed world. Harvard economist Dani Rodrik found that there is an inverse-U shaped relationship between both manufacturing and GDP per capita of a given country. Manufacturing employment and GDP share increase as a country gets richer, peak at a certain level, then decline. What’s tough, however, is that manufacturing employment peaks much earlier than GDP share. Workers lose out on the high pay and stability of manufacturing jobs even as manufacturing output increases. The economy continues to grow, but fewer and fewer workers get to share in this growth. In the U.S., manufacturing employment’s peak occurred in 1944, at 39% of nonfarm employment. These days, that figure is 8.8%, and falling still.
Rodrik also found that low-skilled jobs are the first to be lost. Low-skill employment has been decreasing at an increasing rate since the 1990s, while mid-skill employment only started to suffer in 2008.[1] In fact, nearly the entire reduction in manufacturing employment from 1995 to 2009 has come from the low-skill group. Meanwhile, high-skill employment has seen gains.
This trend is troubling, because industrialization once served as the engine of inclusive American growth for nearly a century. Subsequent deindustrialization has ushered in a new era of income inequality and class divergence.
Inseparable from this issue is unionism and the labor movement. Unionization rates peaked around the same time as manufacturing employment – 35% of nonagricultural workers were in a union in 1945. They have declined together ever since, with unionization rates today sitting at 10.7% of wage and salary workers. In the golden age of work in the U.S., both union and manufacturing workers had job stability, work safety measures, wage premiums, and superior benefits, allowing them to support a family with their own income.
As unions and manufacturing declined, low skill workers have had to look somewhere else in the economy to find good jobs. These other work prospects are in sectors that feature slower productivity growth, which translates to largely stagnant wages for workers. The primary result of deindustrialization, whatever its cause, has been an erosion of labor market prospects for low skilled workers.
Until the recent election, economists had a unanimous stance on trade: benefits to American consumers via price reductions and capital inflows outweigh the costs of falling wages and the dislocation of low-skill workers. Policymakers believed that the U.S. labor market was fluid enough for dislocated workers to find new jobs at a similar wage in a different locality. In other words, opening up trade borders would hurt wages for low-skill workers generally, but it wouldn’t target some workers over others.
New work by MIT Economist David Autor, along with his colleagues Dorn and Hanson, shows that these assumptions were wrong – trade does have localized impacts, and our labor market features much more friction than previously thought. His paper finds a connection between a locality’s exposure to the “China shock” – or the rapid expansion of U.S. imports from China starting in the mid-1990s – to localized reductions in U.S. manufacturing employment, labor force participation, wages, salaries, and government transfers, as well as increases in unemployment. In other words, areas that saw increased Chinese imports per capita experienced a localized recession lasting for years. These findings run contrary to the theory that drive free trade policies, which posit that due to labor market flexibility, trade shocks diffuse over the entire nation, manifesting only in slight wage reductions for low-skill jobs.
As expected, neoliberal economists have rushed to defend the status quo. There have since been challenges to Autor’s conclusions, and their validity does remain in question. But even if the empirical work isn’t squeaky clean, their paper has hit on a kernel of truth that cannot be explained away, much less ignored.
There is no denying that our working age population is suffering due to deindustrialization, in which trade liberalization plays a definitive role. Many rural regions have been allowed to decay with little to no economic assistance. At the same time, unskilled workers across the country are being crunched by increasingly poor labor market prospects. Incomes at the bottom of the distribution have been stagnant for decades. While the Fed normalizes monetary policy with the conviction that we are nearing full employment, a sizable chunk of the decline in labor force participation remains unaccounted for by demographic or cyclical explanations. Nearly half of prime age men not in the labor force take pain medication on a daily basis, while all of them report very low levels of emotional well-being. Opioid addiction rates and deaths from overdose are skyrocketing. This is the true American Carnage.
If we can take anything from Trump’s message and how it resounded with voters, it’s that Americans want good jobs, which are nothing compared to the slight extensions of the welfare state pushed by the center-left. And voters are correct in wanting good jobs over an expanded safety net. America’s true welfare system is less easily seen, because it’s predicated on Americans having well-paying jobs. Hundreds of billions of dollars per year are doled out in the form of tax credits for employer-provided healthcare, mortgage interest, retirement savings, childcare, and more. As tax credits, these benefits increase with an individual’s earnings. When workers are shut out of stable, well paying jobs, they are shut out of access to these benefits – left only with a stagnating wage, which increases barriers to social mobility.
The path forward for progressives is to reach out to the low-skill, rural workers that were enticed by Trump’s promises of jobs. These people aren’t irredeemable, even if they voted Trump in order to maintain our nation’s racial hierarchy. Americans care more about having a job than to be surrounded by people who look like them. Misguided convictions about the criminality or laziness of immigrants and minorities will evaporate when voters have confidence in the politicians who are advocating for their inclusion.
In practice, this starts with vigorous communication and organizing, with the message that Trump and Congressional Republicans are acting to make lives worse for their voters. It will require a hard pivot on trade, including talk of spending packages to revitalize decaying rural localities. There is plenty of social space for a revitalized labor movement, this time based around low skill workers, to fight against the increasing profits, offshoring, tax dodging, wage suppression, blockbuster mergers, and financial deregulation of the corporate world. Somehow, someway, the Democratic Party needs to push policies that benefit the employment prospects of low-skill workers. No more neoliberalism, no more weak expansions of the social safety net; rather, infrastructure development, investment in job training and education for high-skill manufacturing, sector partnerships, and the re-shoring of industrial work. In other words, active labor market policies. Jobs matter more than anything else.
[1] This relative abolition of low-skill industrial labor has likely contributed to the breakdown in the labor productivity-wages link, as manufacturing wages were the highest and most productive at the time the two were closely linked. The breakdown itself mirrors the decline in low skill manufacturing employment. As low-skill jobs were destroyed, non-manufacturing employers could increasingly get away with paying less and less due to reduced competition and the excess low-skill labor supply.
Deindustrialization is a well-documented trend in the developed world. Harvard economist Dani Rodrik found that there is an inverse-U shaped relationship between both manufacturing and GDP per capita of a given country. Manufacturing employment and GDP share increase as a country gets richer, peak at a certain level, then decline. What’s tough, however, is that manufacturing employment peaks much earlier than GDP share. Workers lose out on the high pay and stability of manufacturing jobs even as manufacturing output increases. The economy continues to grow, but fewer and fewer workers get to share in this growth. In the U.S., manufacturing employment’s peak occurred in 1944, at 39% of nonfarm employment. These days, that figure is 8.8%, and falling still.
Rodrik also found that low-skilled jobs are the first to be lost. Low-skill employment has been decreasing at an increasing rate since the 1990s, while mid-skill employment only started to suffer in 2008.[1] In fact, nearly the entire reduction in manufacturing employment from 1995 to 2009 has come from the low-skill group. Meanwhile, high-skill employment has seen gains.
This trend is troubling, because industrialization once served as the engine of inclusive American growth for nearly a century. Subsequent deindustrialization has ushered in a new era of income inequality and class divergence.
Inseparable from this issue is unionism and the labor movement. Unionization rates peaked around the same time as manufacturing employment – 35% of nonagricultural workers were in a union in 1945. They have declined together ever since, with unionization rates today sitting at 10.7% of wage and salary workers. In the golden age of work in the U.S., both union and manufacturing workers had job stability, work safety measures, wage premiums, and superior benefits, allowing them to support a family with their own income.
As unions and manufacturing declined, low skill workers have had to look somewhere else in the economy to find good jobs. These other work prospects are in sectors that feature slower productivity growth, which translates to largely stagnant wages for workers. The primary result of deindustrialization, whatever its cause, has been an erosion of labor market prospects for low skilled workers.
Until the recent election, economists had a unanimous stance on trade: benefits to American consumers via price reductions and capital inflows outweigh the costs of falling wages and the dislocation of low-skill workers. Policymakers believed that the U.S. labor market was fluid enough for dislocated workers to find new jobs at a similar wage in a different locality. In other words, opening up trade borders would hurt wages for low-skill workers generally, but it wouldn’t target some workers over others.
New work by MIT Economist David Autor, along with his colleagues Dorn and Hanson, shows that these assumptions were wrong – trade does have localized impacts, and our labor market features much more friction than previously thought. His paper finds a connection between a locality’s exposure to the “China shock” – or the rapid expansion of U.S. imports from China starting in the mid-1990s – to localized reductions in U.S. manufacturing employment, labor force participation, wages, salaries, and government transfers, as well as increases in unemployment. In other words, areas that saw increased Chinese imports per capita experienced a localized recession lasting for years. These findings run contrary to the theory that drive free trade policies, which posit that due to labor market flexibility, trade shocks diffuse over the entire nation, manifesting only in slight wage reductions for low-skill jobs.
As expected, neoliberal economists have rushed to defend the status quo. There have since been challenges to Autor’s conclusions, and their validity does remain in question. But even if the empirical work isn’t squeaky clean, their paper has hit on a kernel of truth that cannot be explained away, much less ignored.
There is no denying that our working age population is suffering due to deindustrialization, in which trade liberalization plays a definitive role. Many rural regions have been allowed to decay with little to no economic assistance. At the same time, unskilled workers across the country are being crunched by increasingly poor labor market prospects. Incomes at the bottom of the distribution have been stagnant for decades. While the Fed normalizes monetary policy with the conviction that we are nearing full employment, a sizable chunk of the decline in labor force participation remains unaccounted for by demographic or cyclical explanations. Nearly half of prime age men not in the labor force take pain medication on a daily basis, while all of them report very low levels of emotional well-being. Opioid addiction rates and deaths from overdose are skyrocketing. This is the true American Carnage.
If we can take anything from Trump’s message and how it resounded with voters, it’s that Americans want good jobs, which are nothing compared to the slight extensions of the welfare state pushed by the center-left. And voters are correct in wanting good jobs over an expanded safety net. America’s true welfare system is less easily seen, because it’s predicated on Americans having well-paying jobs. Hundreds of billions of dollars per year are doled out in the form of tax credits for employer-provided healthcare, mortgage interest, retirement savings, childcare, and more. As tax credits, these benefits increase with an individual’s earnings. When workers are shut out of stable, well paying jobs, they are shut out of access to these benefits – left only with a stagnating wage, which increases barriers to social mobility.
The path forward for progressives is to reach out to the low-skill, rural workers that were enticed by Trump’s promises of jobs. These people aren’t irredeemable, even if they voted Trump in order to maintain our nation’s racial hierarchy. Americans care more about having a job than to be surrounded by people who look like them. Misguided convictions about the criminality or laziness of immigrants and minorities will evaporate when voters have confidence in the politicians who are advocating for their inclusion.
In practice, this starts with vigorous communication and organizing, with the message that Trump and Congressional Republicans are acting to make lives worse for their voters. It will require a hard pivot on trade, including talk of spending packages to revitalize decaying rural localities. There is plenty of social space for a revitalized labor movement, this time based around low skill workers, to fight against the increasing profits, offshoring, tax dodging, wage suppression, blockbuster mergers, and financial deregulation of the corporate world. Somehow, someway, the Democratic Party needs to push policies that benefit the employment prospects of low-skill workers. No more neoliberalism, no more weak expansions of the social safety net; rather, infrastructure development, investment in job training and education for high-skill manufacturing, sector partnerships, and the re-shoring of industrial work. In other words, active labor market policies. Jobs matter more than anything else.
[1] This relative abolition of low-skill industrial labor has likely contributed to the breakdown in the labor productivity-wages link, as manufacturing wages were the highest and most productive at the time the two were closely linked. The breakdown itself mirrors the decline in low skill manufacturing employment. As low-skill jobs were destroyed, non-manufacturing employers could increasingly get away with paying less and less due to reduced competition and the excess low-skill labor supply.