jeff e. biddle is professor of economics at Michigan State University. jeff e. biddle
Illustrations by dan hamermesh is Sue Killam professor emeritus at the University of Texas at Austin and a research fellow at the Institute for Labor Economics in Germany.
Published January 24, 2023
With unemployment near record lows and reportedly millions of disaffected postpandemic workers wondering whether they must return to the daily commute, labor activists are pushing radical changes in work hours. And they recently got a boost from reports that a pilot program cutting days and hours without cutting pay in Britain has gotten surprisingly little pushback from employers.
Back in the United States, the discontent is reflected in the California legislature’s first credible initiative to mandate a four-day, 32-hour week, with a 50 percent wage penalty for overtime. This would be a major change from the current federal mandate of 50 percent overtime pay after 40 hours, a provision that has been essentially unchanged since the landmark Fair Labor Standards Act was passed in 1938. And while it seems to have stalled, the idea is hardly dead.
Stepping back, it’s important to note that interest in the four-day workweek conflates two separate issues: (a) a reduction of total working time, and (b) the squeezing of current work hours into fewer days each week. These would have substantially different effects, both on earners’ living standards and on how they spend their non-work time. So to understand the potential impact of a four-day week — or just to make sure people are not conflating separate issues — we lay out where the U.S. has been on the issues and where we might go with changes driven by market forces or government intervention.
A Two-Minute History
It’s easy to look up how average annual working time for American workers has changed over the past 40 years. By the same token, data from the OECD offer a handle on how time actually on the job (excluding vacations and holidays) differs across industrialized countries now and how it has changed over the decades.
In the 1970s, the United States was in the middle of the pack in terms of average hours. The numbers show that the stereotype of the overburdened Japanese worker was, in fact, accurate in 1979 — the Japanese worked much more than their American counterparts. Not today, though: Americans who are in the labor force (this is an important distinction) average more hours per year than their counterparts in every other rich country. (We are about in the middle in terms of the fraction of adults who participate in the labor force — that is, who perform any paid work at all.)
Those on the left generally attribute our long hours to a capitalist-inspired ethos of shop-’til-you-drop, with bombardments of advertising spurring us to work more to be able to afford the 83-inch OLED flat screens and three-camera smartphones we are beguiled into purchasing. However, if you have spent much time in Europe or Japan, you know that the assault of advertising there was — and remains — as pervasive as it is in the United States.
So why is the U.S. the workaholic outlier? Other countries have mandated long paid vacations that clip a lot of hours from the work year. In the U.S., paid vacations elude almost half the workforce, and it is unusual for workers who do have paid vacations to get more than three weeks off. Also, in most other rich countries there are many more government-mandated paid holidays than in the United States.
Those countries created shorter-hour mandates through the political process, responding to pressures from workers and to the willingness of employers to acquiesce if their competitors were similarly affected. Not so in the United States.
What Might Be Done
If exceptionally long hours are seen as a problem — and we realize that a lot of people have that view — what could be done about it? Mandating an eight-hour, four-day workweek would do the trick. However, evidence from California’s long-running experience with a 50 percent wage premium after 40 weekly hours and eight hours each day suggests that imposing this mandate on work hours beyond 32 would not greatly reduce actual time on the job. Employers are responsive to labor costs — but not to the extent that a 50-percent penalty exacted only on the extra hours of work would lead them to cut weekly hours by eight to avoid paying mandated overtime penalties. An educated guess based on the impact of California’s existing daily overtime penalty is that weekly work time would decrease by two hours at most. Remember, moreover, that even California’s liberal legislature hesitated in imposing a disincentive this stringent after 32 weekly hours on the job.
What about the British demonstration suggesting that once employers tried it, they’d buy in voluntarily because it was truly a free lunch — that workers got more done working fewer hours? We think something was lost in translation.
The happy headlines about higher productivity implied that UK workers’ fifth eighthour day on the job actually reduced their total weekly output, and that if hours are cut workers will be so well-rested and inspired that their output in a shorter week will exceed what they had produced in the longer workweek. But if this were correct, employers ought to have long ago discovered that paying their employees to show up for the fifth day got them nothing — actually, less than nothing. Or, in econspeak, it implies that the marginal product of labor after 32 hours a week is negative.
Now, we are sure we do not live on Bizarro World, the cube-shaped planet imagined by DC Comics in 1960 where everything works backwards. Hence there’s little doubt that any broad initiative cutting average weekly hours from 40 to 32 would lower output. How much lower is hard to say, although evidence from the scientific management era of elaborate time and motion studies in the 1920s is suggestive. Probably not 20 percent lower, since it is likely that the last eight hours per week, be they during the fifth day of work or during the last 96 minutes each day, are of less value to employers — that is, the marginal product of labor diminishes (but remains positive) with more time spent on the job.
In the United States’ mostly competitive economy, the reduced output would lower GDP and almost certainly lead to lower total earnings. Thus, in thinking about cutting worktime, be it by reducing days per week, hours per day or weeks per year, we need to recognize the trade-off: less total purchasing power, but more time to enjoy our (albeit reduced) pay.
Back to Reality
It is, of course, possible to alter the number of days worked without changing hours per week or weeks per year by changing the length of the workday. Indeed, the number of people working four-day weeks has been increasing in the U.S. for at least a half century without government mandates. While the data are only sporadically available for the United States, they do show that just 3.5 percent of all American employees were working four days per week in 1973 compared to 8.5 percent today. Even among full-time workers (defining full-time as 30+ weekly hours) the number of employees working four-day weeks rose from 1.9 percent in 1973 to 6.6 percent in 2018 — an increase from a bit over 1 million American workers in the early 1970s to over 9 million in the late 2010s.
This pretty big change has been occurring without government carrots or sticks. Keep in mind, too, that neither the average workweek nor average annual work hours have declined over this period. In 2018, among those fulltime American employees who toiled on a four-day schedule, three were working 40 or more weekly hours for each one working just 30-32 hours. That 3:1 ratio prevailed back in 1985, as well.
The incidence of four-day workweeks varies tremendously across industries. Among full-time workers in 2018 it was still rare (less than 5 percent) in finance, mining and construction. It was much more common (more than 10 percent of workers) in health care, food service, transportation and among first responders. Even within industries, the prevalence of four-day work differs across demographic groups. Looking only at full-timers, four-day weeks are less common among older workers and more educated workers, as well as among Black and Latino workers, and women. (However, as might be expected, the four-day week is relatively common among women who have pre-school children at home.)
One might guess that the increase in the four-day workweek arose because the composition of the workforce has changed greatly over the half-century. Its composition certainly did change: the full-time workforce increased from 36 to 45 percent female, while white non-Hispanic representation fell from 84 to 64 percent; and the numbers with at least some college rose from 14 to 42 percent. But this doesn’t get to the heart of the matter because the increase in four-day work has occurred to varying extents within each demographic group.
It is also tempting to explain the gradual market-driven rise in four-day work by the vast changes in the structure of American industry. During the past half-century, workers in service industries increased from 19 to 32 percent of full-time employment, while the manufacturing labor force fell from 25 to 14 percent. And it would make sense that services — everything from warehouse workers to university professors — can be organized more flexibly than, say, autoworkers. Yet, in fact, the four-day workweek increased in nearly all industries. The rise of four-day work appears to have been linked, on the one hand, to managerial and technical trends that allow employers more flexibility in labor policies and, on the other, to employees’ increasing interest in stretching their days off while still being paid for full-time work.
Germany and the Netherlands are among the few other industrialized countries for which similar information on days worked per week can be calculated over long periods. Dutch data are available in five-year intervals from 1975 through 2005. Partly because the Dutch work fewer hours per year and partly because of the way the data are collected, the percentage of full-time workers on four-day weeks is much higher than in the U.S. But, as in the U.S., this percentage has risen fairly steadily over the past half-century.
The story is similar in Germany. Since 1992, Germany has occasionally amassed data on days worked per week, as well as collecting the usual information on hours worked per week. Defining full-time work as 30+ weekly hours, the German numbers show the same upward trend as in the U.S. and the Netherlands. However, the percentage of four-day, full-time workers in Germany is less than in the Netherlands or America, perhaps because of the more rigid culture of work.
In both the U.S. and the Netherlands, the overwhelming portion of four-day workers are on the job on four consecutive days and are free for three consecutive days, allowing mini-vacations. Note, too, that four-day work saves some commuting costs — both in time and money — although U.S. evidence suggests that this is apparently not a significant motive because the four-day week does not seem to be more common in traffic-challenged cities. And, of course, employers have a big influence here: it may well have become easier for them to sustain profits while giving workers more flexibility in days worked.
Four-day weeks need not imply fewer hours of work, but they do imply differences in the timing of when people work for pay. With this different timing, and with no change in the length of the workweek, would workers need to accept a pay cut to convince employers to accommodate them?
The answer depends on what amounts to a race between workers’ preferences and employers’ technologies: if enough employers can have at least some of their workers on four-day schedules at no cost or even with some gain in productivity, and if sufficiently small numbers of workers are willing to forego earnings in order to get their leisure bunched, workers on four-day schedules might actually earn more than those on more standard schedules.
In the U.S., it turns out that full-time workers on four-day schedules earn less per hour than their counterparts in the same broadly defined industries who are on fiveday schedules. But only a bit less: in the past 25 years, four-day workers have been sacrificing only 1 percent of their potential earnings in order to have their employers accommodate their desires for bunched leisure. Apparently, employers’ willingness/ability to offer four-day work is nearly in balance with workers’ willingness to give up a bit of their pay to enjoy more consecutive days off.
This has not always been the case, though. In the previous quarter-century the much smaller number of four-day workers received 4 percent less per hour than comparable fiveday workers. The ease with which employers can offer this alternative schedule has apparently grown more rapidly than have workers’ desires to work only four days. And it follows that the trend toward concentrating 40 hours of work into four days now costs the American economy very little.
Outside their work hours, four-day workers who sustain full-time workloads behave very much like other workers with the same demographic profiles. On an average day in any week, four-day workers do watch about 5 percent less television. But in other respects, their days off are filled with the same activities in roughly the same proportion. Like all workers, five-day workers sleep more and do more housework on days when they are not on the job.
Where is all this heading? That is an especially hard question to answer, both in light of always-changing workplace technologies and the tremendous shocks to the timing and location of work produced by Covid-19 that may or may not fade along with the pandemic. But a few things are pretty clear. If commuting costs — especially if the “time costs” of commuting — drop because of the growth of telecommuting, the incentive for four-day workweeks will decline, and we might see little further growth in four-day work. That seems especially likely knowing that people generally prefer variety in their lives, and that higherincome people do more different things with their time each day. Increasing incomes will encourage more variety in our daily life, and we will see more variety in how people spend their time, including perhaps more mixing of work and leisure on the same day.
More important, what is going to happen to the number of hours Americans work by the week and year — which, except for small cyclical fluctuations, has remained remarkably unchanged for quite a while? Of course, we don’t know for sure. But it is striking that with all the changes in the U.S. economy over the past 75 years — changes in the proportion of two-earner households, the reduction in jobs requiring heavy manual labor and substantial increases in earnings for some skill groups — there has been such little change in total hours of toil. And to us, that suggests it is unlikely that market forces alone will lead to shorter workweeks in the foreseeable future. Maybe bigger flatscreens do trump more time with the family.
What’s Big Brother to Do?
These considerations raise two questions of public policy. First, should governments encourage four-day work? The evidence suggests that simply creating strong financial incentives for business to shorten workweeks — as France did late in the 20th century — will not produce a sea change in the timing or amount of work. So unless governments mandate three-day weekends, we are not likely to see more rapid change. We doubt, though, that any industrialized country will impose such a mandate.
If enough workers wish to work less and are willing to sacrifice some of what they would otherwise earn, the change will eventually be reflected in a combination of market outcomes and legislated mandates.
Second, and more important, should governments encourage fewer hours of work across the year? They could follow the lead of the Europeans, who mandated reductions in hours between the end of World War II and the 1990s. As a practical matter, following suit in the U.S. would require mandating longer (and paid) vacations, and increasing the rate of “holiday creep” — the continuing expansion of the number of national holidays.
But here, while we acknowledge that unregulated labor markets don’t always efficiently reflect workers’ and managers’ preferences, we think that holidays, too, should be a matter of private choice. If enough workers wish to work less and are willing to sacrifice some of what they would otherwise earn, the change will eventually be reflected in a combination of market outcomes and legislated mandates.
At the risk of repeating ourselves, remember that in a relatively competitive labor market (like ours), less work means lower material living standards.
Now, say, 10 percent fewer hours would probably translate into less than 10 percent less pay because the last 10 percent of the time on the job is typically less productive than the first 90 percent. And if Americans do opt for more leisure at the cost of less income, we have no problem with that. Our concern, though, is that advocates for shorter hours are inclined to gloss over the reality that this lunch is not free, that we must choose between more leisure and more spending power.