Labor unions aren’t “booming.” They’re dying.

Every once in a while, reporters see a few successful unionization drives in the US, like at Starbucks or Amazon, and conclude that the US is in the midst of a labor union resurgence, that unions are “booming,” or that they’re “suddenly and rapidly rebounding.”

I am, like all non-management Vox staff, a Writers Guild of America, East member, and a former member of Vox’s union bargaining committee, so I empathize with the urge to be optimistic about the future of organized labor, especially at a time when my comrades in the film and TV industries are striking for a better contract. I stand in solidarity with them and hope they get an excellent contract.

But I think it’s even more important to be honest about the situation. Organized labor is not booming, rebounding, or in a resurgence of any kind. Instead, it is in decline, as it has been for many years.

Official data from the Bureau of Labor Statistics starts in 1983. That year, 20.1 percent of all workers were in a union. That’s down to 10.1 percent as of 2022 — the lowest it’s ever been in that time frame. The decline has been basically continuous, with brief interruptions in 2008 and 2020 as non-unionized workers lost jobs faster than those with union protections. While public-sector unionization has fluctuated a bit (it fell from 36.7 percent to 33.1 percent from 1983 to 2022), by far the sharper decline is in the private sector, where rates fell from 16.8 to 6 percent.

Planet Money’s Greg Rosalsky put it well in a piece earlier this year: “While there was an uptick in labor organizing in 2022, we’re hardly witnessing a rejuvenated movement strong enough to dramatically reverse unions’ long-run decline.”

Starting the data at 1983 gives a misleading picture: The decline of unions began well before then.

Harvard economist Richard B. Freeman has put together data on union membership going back over 100 years. It shows that the share of households with a union member was around 10-11 percent going into the Great Depression. Starting in 1937 (not coincidentally the year the Supreme Court upheld the pro-union National Labor Relations Act passed two years earlier), you see a dramatic rise in membership. The rate went from 13.2 percent in 1936 to 26.6 percent in 1938. The rate peaked at around a third of households, and stayed in that range for decades. But by the mid-1950s, a slight but perceptible decline was already starting, which has continued ever since.

I’ve seen two major theories for why this happened. The first emphasizes politics: Countries with more left-wing governments have seen smaller declines in unions. In Canada, for instance, the share of workers in a union has fallen, but the fall is less stark than in the US, which might be explainable by its more pro-union laws.

The second emphasizes the fact that union firms tend to expand their workforces less quickly than other firms. That makes sense: Unions raise wages, so union workforces cost more. But over time, this effect means a greater and greater share of the workforce is non-unionized because non-unionized firms are able to grow faster.

In a landmark 2001 paper, economist Henry Farber and sociologist Bruce Western credited this as a major factor behind union decline in the US. They estimated that unions would have to increase their organizing rate sixfold just to keep the US membership rate constant. To increase unionization, they’d need an even more dramatic and improbable explosion in organizing.

I lean toward the latter theory. It helps explain why you didn’t see a collapse in union rates when hostile governments (like those of Ronald Reagan or George W. Bush) came to power, but instead the same gradual decline as occurred under Democrats.

Whatever explanation you choose, any attempt at union revitalization will require much more than organizing a few Starbucks locations. It will require wholesale change to labor law.

The political scientist David Madland’s book Re-Union gets into the details well, but the gist is you need to find ways to organize unions across whole sectors, not just workplace by workplace. In many European countries, firms don’t pay a penalty for paying good union wages; union contracts are “extended” to whole sectors. If UPS drivers win a good contract, FedEx would then have to abide by those terms too, even though it doesn’t have a staff union.

This would be an ambitious change. The PRO Act, the labor movements’ big priority in Congress (which is currently dead in the water given Republican control of the House), wouldn’t do much to further it; that act would mostly strengthen the existing workplace-based system, which is valuable but insufficient. Some states like California are experimenting with sectoral bargaining, but we’re in very, very early days yet.

The future of labor, I think, lies much more heavily in legal reform efforts meant to enable that kind of broader bargaining than it does in a few heavily publicized elections at individual companies. It’s not sexy work, but it’s the only thing that could return organized labor to the power it once had in the US.