[ AD HAS BEEN REMOVED FROM THIS STORY ]
It is well established that unions provide benefits to workers — that they raise wages for their members (and even for nonmembers). They can help reduce inequality.
A new study suggests that unions may also help children move up the economic ladder.
Researchers at Harvard, Wellesley and the Center for American Progress, a liberal think tank, released a paper Wednesday showing that children born to low-income families typically ascend to higher incomes in metropolitan areas where union membership is higher.
The size of the effect is small, but there aren’t many other factors that are as strongly correlated with mobility. Raj Chetty of Stanford and Nathaniel Hendren of Harvard, who pioneered this method of examining economic mobility, established five factors that are strongly correlated with a low-income child’s likelihood of making it into the middle class: the rate of single motherhood in an area, the degree of inequality, the high school dropout rate, the degree of residential segregation and the amount of social capital, as measured by indicators like voter turnout and participation in community organizations.
Single motherhood is the most strongly correlated factor with mobility. The latest study, which relied on the Chetty/Hendren data, says union membership is roughly as strongly correlated with mobility as the other four factors.
"It’s a striking relationship," said Lawrence Summers, the former Treasury secretary and Obama economic adviser, who is participating in a discussion with some of the study’s authors Wednesday. "It’s further grounds for concern about the decline of unionism in the United States."
The authors posit a variety of reasons for why higher rates of unionization tend to coincide with greater mobility, beyond the effect on parents’ wages, which would seem to be the most obvious way unionization could matter.
Their most interesting explanation is that unions are effective at pushing the political system to deliver policies — like a higher minimum wage and greater spending on schools and other government programs — that broadly benefit workers. Perhaps not surprisingly, three cities that appear to reflect the union effect — San Francisco, Seattle and New York — are all jurisdictions where the minimum wage is rising substantially (although for New York it is only for workers in fast-food chains.).
The researchers looked at the expected income of people ages 29 to 32 whose parents were at the 25th percentile of income nationally when they were teenagers. They found that a 10-percentage-point increase in the rate of unionization in an area coincided with a rise of an additional 1.3 points on the income distribution as the average child becomes an adult.
Let’s take the example of the average metro area where about 16 percent of workers were unionized, and children whose parents were in the 25th percentile of income earners nationally ended up at the 40.7 percentile on average as adults. A simple application of the author’s finding implies that, in a metro area where 26 percent of workers were unionized, the average child from the same place in the income ladder would end up in the 42nd percentile.
The correlation remains statistically significant even when the researchers controlled for a variety of other social and economic variables, like the child poverty rate and median house value.
"I would have thought we could have found things that might have killed off the effects," said Richard B. Freeman, a labor economist at Harvard who was one of the study’s authors. "And we basically didn’t."
Moreover, the benefits aren’t exclusive to low-income children. The researchers also find that a 10-percentage-point increase in the rate of union membership is associated with a 3 percent to 4.5 percent increase in the incomes of all children — regardless of their parents’ income. (The differences in the 3 percent and 4.5 percent arise from the number of demographic variables the researchers control for.)
Because there is more upward mobility in areas with greater unionization generally, the researchers concluded that the result was not because the children of union members seized opportunities from other children.
It’s important to emphasize that the study does not establish causality — the authors can’t prove that unions are driving the improvement in mobility. For that matter, they don’t attempt to. The finding establishes only that, in their words, "mobility thrives in areas where unions thrive."
As Scott Winship, a fellow at the conservative Manhattan Institute points out, this is not a minor limitation. It’s entirely possible that unobserved features of a given metro area cause both the increase in unionization and the increase in mobility.
Winship notes, for example, that both unionization and mobility tend to be relatively low in Southern cities like Atlanta and Charlotte, N.C. "There may be a regional difference" that explains both, he said.
To help pin down causality, he said, it would be helpful to look at variations across geographically similar areas, like cities that have different rates of unionization and variations in mobility on either side of a state border.
Still, the result is especially telling given a common critique of unions — that they may raise wages for workers in an area, but they lower employment by making marginal workers unaffordable. Even if that’s the case, the current study suggests that the benefits of greater unionization are outweighing the costs: Children are doing better on average when unionization for their parents’ generation is higher, even if the higher rates of unionization could theoretically lower employment.
If unions were doing more harm than good, we wouldn’t expect to see mobility rise as the rate of unionization does. (For what it’s worth, Freeman notes, most studies find that unions have little negative effect on employment.)
An intriguing question is whether there might be something unique about the way unions affect mobility, independent of their effect on inequality or dropout rates. Is it possible that there’s a union special sauce? The authors don’t grapple with this question directly, but buried deep in the study is a potential clue.
It comes when the authors use a second, more detailed data set to analyze whether parents’ union membership tends to increase the wages of their children. (The second data set pairs specific individuals with their parents, rather than relying on metro-area averages, and controls for a variety of demographic characteristics that might also affect wages — like race, ethnicity, marital status and education — so they can isolate the effect of union membership.)
The authors find that children with fathers who belong to a union have significantly higher wages than children who don’t. But when it’s the mother who belongs to a union, only the wages of daughters rise.
What might be going on here? It’s possible that the explanation is sociological: Daughters with a mother who belongs to a union may be more likely to work themselves, which means they’re more likely to have higher wages. Or, put differently, union membership is helping to change social norms.
"I like to think it’s the role model effect," said Brendan V. Duke, another one of the study’s authors at the Center for American Progress, who concedes that the explanation is speculative. (Eunice Han of Wellesley and David Madland of the Center for American Progress were the other authors.)
And that, in turn, suggests something potentially important, though equally speculative, about the effects of unions more broadly: Higher rates of unionization may give rise to certain norms that instill a greater sense of agency in workers.
For example, people who belong to unions are generally aware that they have certain rights in the workplace and are encouraged to speak up if they believe they’ve been mistreated. It’s the kind of norm that could leach out into a broader population — to both union members and their nonunion peers — if unions are sufficiently visible and active, which could in turn help boost economic mobility.
Freeman believes there may be something to this, but notes that the study did not explicitly pursue this line of inquiry. "I’m thinking of which student might we get to do an undergraduate thesis on this," he said.
Whatever the possible mechanism, the study highlights the potential of policies and institutions that are important both to the individuals directly affected by them and to those affected only indirectly. Union membership can lead to a virtuous cycle, Summers asserts, improving outcomes for union members who then positively affect their peers, who then positively affect the union members, who then positively affect their peers, and on and on.
"When you work all that out, things that have a small effect at the individual level can have a larger aggregate effect," he said. "Freeman et al have demonstrated that one of those things is the incidence of unionism."