The decades-long decline of unions in the United States is a long-studied phenomenon, but now several academics are wondering whether past research underestimated the role of the decline of unions in the rise of economic inequality. The new research also raises the ancillary question of how could unions rise again to prominence to help reduce inequality.

A variety of research shows that the decline in unionization was a significant factor in the rise in income inequality in high-income countries. A recent paper by sociologists Bruce Western of Harvard University and Jake Rosenfeld of the University of Washington is widely cited. Their research finds that about one-fifth to one-third of the increase in wage inequality can be explained by declining union membership.

Similarly, a recent essay by International Monetary Fund economists Florence Jaumotte and Carolina Osorio Buitron previews their new research that finds the decline in unionization also affects top-level inequality in the industrialized economies. According to Jaumotte and Buitron, “the decline in unionization explains about half of the 5 percentage point rise in the top 10 percent income share” in these nations.

But Adam Ozimek, an economist at Moody’s Analytics, makes a very important observation. How exactly will unions return to prominence? Ozimek cites a variety of research that shows employment growth at unionized firms is slower than employment at nonunionized firms. Specifically he cites a paper by Harvard University economist Richard Freeman that makes this point—for the unionization rate to increase, employment at unionized firms would have to increase faster than employment at non-unionized firms.

So what could make employment at unionized firms grow relatively faster compared to non-union firms? The results Freeman cites would have us believe that such an outcome is unlikely. A steady increase in the unionization rate seems unlikely.

Yet in the same paper, Freeman notes that increases in unionization rates in the past have come from large and sudden increases in union membership. So a sudden reform of labor law might make unionization much easier. Or a sharp increase in demand for unionization might happen. Or a combination of both—a more likely scenario—could happen.

In the mid-20th century, labor union were an integral part of the U.S. economy. More than 30 percent of private-sector workers were members of unions and some of the most successful industries of the time, namely the auto industry, were highly unionized. Early in the 21st century, the status of unions isn’t so strong. Less than 7 percent of private-sector workers are members of unions and once highly unionized sectors have seen their heydays.

But one advantage that unions do have moving forward is that the growing industries in the U.S. economy are seemingly less likely to move abroad in the same way heavily unionized manufacturing jobs did over the past several decades. In this regard, service-sector jobs may be more likely to stick around.

The take-away is this: Increased unionization in the United States would almost certainly reduce wage and income inequality. But the pathway to a higher rate is steep. A sudden jump seems the only way up. But what will cause or even allow such an increase? That remains to be seen.