There’s something curious about the labor force in the United States. Identical jobs and industries have become unionized in some states while remaining nonunionized in others. Unionization levels vary greatly from state to state.
As of 2013, 20 to 25 percent of employed residents in some states belonged to a union, while in other states the union density was only 3 to 5 percent.
"So the people who had jobs said, ‘Let’s unionize to make sure these farmers don’t take our jobs’"
According to the Bureau of Labor Statistics, in 2015 roughly half of the nation’s 14.8 million union members lived in just seven states: California, New York, Illinois, Pennsylvania, Michigan, Ohio, and New Jersey. Yet these states accounted for only one-third of paid employment nationally. And those same states have held the highest unionization rates for decades. So why do some states remain heavily unionized while others do not? “It turns out there was something that happened in the 1930s that set the rank of unionization in place across states in the United States, and that rank has stayed roughly the same ever since,” says Lauren H. Cohen, the L.E. Simmons Professor of Business Administration at Harvard Business School.
The thing that happened was the Dust Bowl: a series of severe dust storms and droughts that decimated farms in the Great Plains during the 1930s, forcing thousands of families to abandon their property. Many migrated to close-by cities, often in California but also in other states, in hopes of finding jobs.
Alas, the Dust Bowl coincided with the Great Depression. Jobs were scarce, and those who still had jobs were loath to lose them to migrants. And so they unionized.
“Let’s say you were a subsistence farmer,” Cohen explains. “The drought dried up your crops. You still had to feed your family. So you traveled to the closest city and tried to get a job. Of course, that put pressure on people who did have jobs. They were working for a dollar an hour, and you were willing to come and do the same job for 50 cents. So the people who had jobs said, ‘Let’s unionize to make sure these farmers don’t take our jobs.’”
The relationship between the Dust Bowl and present-day unions is detailed in Cohen’s new working paper, The Impact of Forced Migration on Modern Cities: Evidence from 1930s Crop Failures, co-written by HBS finance professor Christopher J. Malloy and Quoc Nguyen, an assistant finance professor at the University of Illinois at Chicago.
Determining what drove unionization
In setting out to explain the differences in unionization rates by location, the researchers noticed that state-to-state union levels differed before and after the 1930s. Whether a state had a percentage of unionized employees in 1929 did not predict whether that state would have a high percentage of unionized employees today. However, unionization density remained consistent, in terms of which states ranked highest and lowest, from 1939 onward.
Importantly, the variations occurred within individual industries—for example, unionization rates within the manufacturing sector have remained high in the same states and low in the same states for the past three-quarters of a century. Thus, the variation in state unionization habits couldn’t be explained by mere differences in a state’s industrial makeup.
In search of additional explanations, the researchers looked for significant historical events in the 1930s that might affect some states but not others.
They considered the National Labor Relations Act of 1935, which guaranteed the rights of private-sector employees to unionize and engage in collective bargaining. The act certainly encouraged labor forces to unionize. But a national act applies to the whole nation, so it didn’t explain state-to-state differences in unionization rates.
The Dust Bowl, on the other hand, was a massive external force that affected only certain geographic areas.
To determine whether the Dust Bowl drove unionization, as opposed to some other geographically specific reason, the researchers compared the major drought years of 1930, 1934, and 1936 with non-drought years, 1931 and 1935, in the same geographical areas. They looked at firm-level unionization data from the US Census, the Center for Research in Security Prices, and Compustat; crop failure and interstate migration data from the US Census; and rainfall data from the National Climatic Data Center (now called the National Centers for Environmental Information).
Overall, they found that migration rates in years with normal rainfall did not predict unionization levels today. But unionization levels increased dramatically, in conjunction with migration, during drought years. And the unionization associated with drought years has had lasting effects, all the way to the present.
The drought-year unionization density predicted relative unionization density in 1943, 1953, 1973, etc., all the way up to 2013. (These findings held true only in those industries in which migrants tried to find jobs during the Dust Bowl—manufacturing jobs, for instance, but not teaching jobs.) In other words, it wasn’t all migration but only the migration related to the Dust Bowl droughts that predicted modern unionization patterns.
“Droughts caused migration. Migration caused pressure on the current workforce. Pressure caused the workforce to unionize. And that unionization just has an incredibly long tail,” Cohen says.
“A one-standard deviation lower drought-year rainfall in an MSA [metropolitan statistical area] predicts unionization rates that are 2.3 percent higher; relative to an unconditional average 2013 unionization rate of 6.9 percent, this implies a 33 percent increase in the MSA-level unionization rate today,” the researchers write.
The takeaway: “What this says is that a completely random event that happened 80 years ago still has a huge impact on the amount of unionization today,” Cohen says.
To Cohen, the findings tell less about the staying power of unions than about the power of random, external forces. In 80 years, nothing has happened in the world of labor unions to shake the effects of some lousy weather.
“I think the reason why this paper is important, especially now, is that unions are a hotly debated issue within policy and within our political process,” Cohen says. “Some people say we absolutely need them. And some say there was a time in history that we needed them, but we’ve outgrown that time. I think both sides have to confront the fact that a fair amount of unionization that exists today was set in place in a random way. So if you want to say unions are great, or if you want to say they’re awful, either way you have to explain why something so obviously great or so obviously awful can be so significantly influenced by something that is essentially random."