Research Highlights Article
November 6, 2024
Origins of the 40-hour workweek
How did workweek restrictions instituted during the Great Depression impact US labor markets?
Source: National Archives and Records Administration, Public domain
In recent years, companies and legislators have begun exploring a four-day workweek, which, if adopted, would be the biggest change since the 40-hour workweek became the norm during the Great Depression.
In July of 1933, the Roosevelt administration introduced the President’s Reemployment Agreement (PRA) to encourage work sharing with the goal of increasing employment overall. Price Fishback, Chris Vickers, and Nicolas L. Ziebarth studied the effects of the PRA in a paper in the American Economic Journal: Macroeconomics and found that employment rose significantly in the month following its adoption.
The PRA allowed companies to show their “business patriotism” by adhering to it in exchange for the right to display the Blue Eagle symbol of the National Recovery Administration. Through radio broadcasts, community gatherings, and door-to-door canvassing, the administration encouraged the nation’s consumers to shop only at firms displaying the Blue Eagle, which made it difficult for firms to say no. The drive resulted in nearly 9 out of 10 establishments signing the PRA within a couple of months of its introduction.
In a lot of industries, this is where the 40-hour workweek comes from. If you look in the micro data, there are plenty of firms with 50- to 60-hour workweeks before the PRA, and that just goes away in a month.
Chris Vickers
In return for a Blue Eagle, the PRA required companies to agree to a maximum workweek of 35 hours. Initially, compliance with the PRA was high, but within months companies ignored the 35-hour limit. Nevertheless, a 40-hour workweek became the norm and was eventually supported by law when the Fair Labor Standards Act was passed in 1938.
“In a lot of industries, this is where the 40-hour workweek comes from,” Vickers told the AEA in an interview. “If you look in the micro data, there are plenty of firms with 50- to 60-hour workweeks before the PRA, and that just goes away in a month.”
The researchers combined two datasets to explore the impact of these changes. The Census of Manufactures in 1933 and 1935 provided information on monthly employment and hours worked across eight industries at the establishment level, and the Bureau of Labor Statistics provided information on monthly earnings from 1933 to 1935 at the industry level across 115 sectors of the economy.
Starting in August 1933, there is clear evidence of establishments “bunching up” at the new workweek limits. For instance, in the steel works industry, the fraction of establishments with a workweek within two hours of the maximum more than doubled between July and September 1933.
Bunching workweek hours
The chart below shows the distribution of workweek hours, scaled by the workweek limits established by the PRA. Bunching around 100 occurs after the PRA goes into effect in July of 1933, as indicated by the shift in the red distributions. No such change occurs before and after July of 1935, as indicated by the blue distributions, when there were no major changes to workweek limits.
Leveraging this bunching effect in a difference-in-differences framework, the authors compared jobs lost above the workweek limit with jobs gained just below the limit, before and after the PRA push.
They found an overall increase in employment in August 1933 of about 24 percent relative to July. These total employment gains persisted through October, but then gradually declined.
However, these estimates only apply to firms with workweeks around the maximum limits. Establishments well below the limits may have also been impacted negatively through indirect equilibrium effects. The authors demonstrate this by constructing a model of the depression-era labor market based on the work of Ben Bernanke, whose model was originally designed to explain rising wages in the presence of mass unemployment. As a result, the bunching estimates are best interpreted as an upper bound on the rise in employment due to the workweek limit.
In tandem with their employment results, the researchers also found that industries directly impacted by the workweek restrictions saw wages grow 9.4 percent faster than their counterparts with shorter workweeks. However, these wage gains were not large enough to stop income from falling overall.
Nevertheless, the authors argue that the Roosevelt administration would likely have viewed the policy as a success.
“I think our work shows that employment-sharing experiments can work. They can generate higher levels of employment,” Vickers said. “While there are downsides, they may not be that severe, at least in the short run. The decline in earnings was not of the same order of magnitude as the increase in employment.”
♦
“Labor Market Effects of Workweek Restrictions: Evidence from the Great Depression” appears in the October 2024 issue of the American Economic Journal: Macroeconomics.