Conservatives and progressives don’t agree on a whole lot, but one thing they do agree on is the need to increase opportunity and mobility. Upward mobility and opportunity are the definition of the American dream. But today, America has less opportunity compared to other advanced economies. A U.S. child born in the bottom 20 percent of the income distribution, for example, has only a 7.5 percent chance of reaching the top 20 percent as an adult. In Denmark, that same child has an 11.7 percent chance of reaching the top 20 percent, while in Canada, it’s a 13.4 percent chance.
If we want to increase mobility here in the United States, we first need to understand why it’s so low. Social scientists are making substantial progress in that direction: A research team lead by Raj Chetty of Stanford University recently used tax records to calculate where in the country a low-income child had the best chance of moving upward. They found that five factors had the strongest geographical relationship—positive or negative—with mobility: single motherhood rates, income inequality, high school dropout rates, social capital, and segregation.
Chetty and his coauthors’ findings have received a great deal of attention: they have been featured in The New York Times several times, and both Hillary Clinton and Jeb Bush have met with Chetty to discuss his research.
But Chetty and his coauthors left one possible driver of mobility unexamined—labor unions. There are several reasons why unions could boost mobility. First, low- or middle-skilled union workers make more than similar non-union workers and could pass those higher earnings onto their children. Second, union jobs may be more stable and more likely to offer health insurance and other benefits that stabilize families’ finances. Third, children who grow in an area with high union membership may be able to join a union when they enter the workforce. Finally, unions are the strongest advocates for middle-class policies that increase mobility—such as spending on education and raising the minimum wage.
But just because a relationship could exist does not mean a relationship does exist. David Madland and I joined with Harvard economist Richard Freeman and Wellesley economist Eunice Han to examine whether there is a link between unions and mobility. Using the same method as Chetty and his coauthors, we examined whether areas with high union membership were also areas with high mobility.
We found a very strong relationship, indeed—a 10 percentage point increase in an area’s union membership is associated with low-income children rising 1.3 percentage points higher in the national income distribution. Mobility’s correlation with unions was at least as strong as its correlation with high school dropout rates and segregation, two widely acknowledged determinants of mobility. Using another measure of mobility, we found that 10 percentage point increase in an area’s union membership was associated with children earning 3 percent higher incomes even after controlling for a host of factors including their parents’ incomes.
But a lot of things could be geographically correlated with mobility without actually causing mobility—one would likely find a negative relationship between mobility and the number of Waffle Houses in a state because mobility is low in the American South. Yet nobody would argue that Waffle House actually causes low mobility. The same could be true for unions.
To gain more confidence in our findings, we used a different dataset that matches adult children with their parents. This allowed us to compare the outcomes of children whose parents were and were not in unions while controlling for their parents’ education, race, age, industry, occupation, and more. We again found a strong relationship—having a union father was linked with 19 percent higher earnings for children. Notably, this effect was much stronger for children whose father did not attend college or had a blue-collar job—that is, unions appear to give opportunity to those who need it most. We also found that children with union parents also attained more education and had better health.
Our findings show that there is a strong relationship across areas and households between unions and intergenerational mobility. Proving whether unions actually cause higher mobility is difficult without performing a randomized control trial, which has become the gold standard in modern empirical economics, but is impractical in this case. The same caveat applies to Chetty and his coauthors’ findings with regards to single mothers and segregation.
The strong union-mobility link should provide a lesson for politicians who have at least rhetorically embraced mobility. A serious policy agenda aimed at boosting intergenerational mobility must include policies that will increase the bargaining power of workers. It is possible that a strong union movement is not simply sufficient for high levels of intergenerational mobility, but it may be necessary. If that is the case, it will be difficult to meaningfully increase intergenerational mobility without also rebuilding unions or some other form of worker voice.
Brendan Duke is a policy analyst at the Center for American Progress.