Are the United Auto Workers justified in seeking higher wages and benefits? Definitely. Is the union overreaching? I don’t know — partly because I’m not an expert on auto industry economics, partly because the actual agreement, when it comes, may be more modest in scope than the union’s current bargaining position.
Still, I think it’s important to understand the context. Both Donald Trump and now Joe Biden promised to revitalize U.S. manufacturing. Trump, with his chaotic trade war, didn’t deliver on his promise. Biden, with his embrace of industrial policy, seems to be having more success, although we don’t yet know how many manufacturing jobs will be created by the surge in manufacturing construction currently underway.
But why should creating more manufacturing jobs be a policy goal? A large part of the answer is the widespread perception that manufacturing jobs are good jobs — jobs that pay well and come with good benefits.
Yet that isn’t necessarily the case. Manufacturing jobs aren’t inherently better than jobs in other sectors. True, there was a long period — basically from Franklin Roosevelt to Ronald Reagan — when manufacturing jobs were, in fact, good jobs. Nostalgia for this period is part of the reason Trump’s slogan “Make America Great Again” got traction with blue-collar voters. But those relatively high wages in manufacturing didn’t just happen: They were negotiated by unions, which were much more powerful in manufacturing than in the rest of the economy. When the power of the unions went away, so did the manufacturing wage premium.
So one way to think about the autoworkers’ strike is that it’s an attempt to make manufacturing good again.
Let’s start with the history. We don’t have all the data we’d like for long-run comparisons — in particular, we don’t have good data outside of manufacturing before the 1960s. But there’s little reason to believe that manufacturing jobs were particularly good jobs before the New Deal.
There were exceptions. In 1914 Henry Ford famously doubled his workers’ wages to $5 a day, which was a lot at the time. His goal was to make boring, repetitive jobs desirable, and hence raise workers’ efficiency, and he seems to have succeeded in doing so. But many, probably most, workers in U.S. manufacturing continued to receive low wages while working long hours.
The big change came in the 1930s and 1940s. Here’s a chart showing real manufacturing wages — the average hourly wage for production workers divided by the Consumer Price Index — from 1919 to 1995:
Notice the steep rise from the late 1930s to the mid-1940s. This corresponds to what Claudia Goldin and Robert Margo call the Great Compression, the sudden reduction in wage inequality that took place mainly during World War II and created the middle-class society I grew up in and that politicians on both the right and the left keep promising to restore.
But simply bringing manufacturing jobs back wouldn’t accomplish that restoration. As you can also see in the chart above, real wages in manufacturing have in general been declining since 1980. The decline has been especially pronounced and persistent in auto manufacturing; here’s a chart that tracks that decline since 1990:
Why did manufacturing jobs get good in the 1940s, then stop being good in recent decades? The answer, almost surely, lies mainly in the rise and decline of unions. Unionization rates surged in the favorable political environment of the New Deal, then withered in the unfavorable environment of post-Reagan America; union power declined even more. Since both unionization rates and union power were stronger at their peak in manufacturing than they were elsewhere, unions were what made manufacturing jobs good.
That diagnosis is upheld by a 2022 paper by researchers at the Federal Reserve Board. Kimberly Bayard, Tomaz Cajner, Vivi Gregorich and Maria D. Tito find that the manufacturing wage premium — the difference between wage rates in manufacturing and elsewhere, once you adjust for differences in worker characteristics — was substantial in the 1980s, but has now disappeared. And they argue that this decline was caused by a sharp drop in unionization, which was initially higher in manufacturing than in other sectors:
This decline in unionization didn’t just reduce the number of workers benefiting from union wages. It also greatly reduced the ability of unions to negotiate higher wages for their workers, so the union wage premium itself fell too:
And the combined effect of fewer unionized workers and a lower union premium, they find, accounts for the bulk of the decline in the premium for manufacturing workers in general.
So what does this say about the current situation? While the Biden administration’s policies may produce a partial revival of manufacturing employment (no, we’re not going back to the days when manufacturing was a third of nonfarm employment), the newly created jobs won’t necessarily be especially good jobs.
This doesn’t mean that a manufacturing revival would achieve nothing: It looks likely that many of the jobs Biden’s polices will create will be in depressed regions of the United States, where any jobs will be welcome. But the new jobs will be good jobs only if the partial revival of manufacturing is accompanied by a restoration of worker bargaining power.
And here’s the thing: There’s no reason a union revival must be restricted to manufacturing. The centrality of manufacturing to the union movement in America was, in a way, a historical accident. Back when favorable political conditions allowed unions to thrive, manufacturing contained most of the big companies you might want to organize. As America increasingly became a service economy, you might have expected unionization to spread to other sectors, as it did in Nordic countries, where two-thirds of workers are still unionized. But America’s shift away from manufacturing — during which Walmart displaced General Motors as the biggest private-sector employer — took place in a political environment that was hostile to union organizing.
So if the autoworkers get major gains as a result of their strike, the lesson you take shouldn’t be that we can make manufacturing jobs good again. It will, instead, be that we can make jobs in general better by helping workers make better bargains.