I have long been a big fan of reduced work time, whether it take the form of more vacations, more time for paid leaves, such as family leave and sick days, or shorter workweeks. The basic logic is that, as our economy gets more productive, we should get some of the benefit in the form of more leisure, instead of just higher income. (Yeah, I know about income inequality, and that most workers have not seen much in terms of higher income in the last half century, but let’s leave that one aside for the moment.)
Anyhow, four-day workweeks always seemed an especially interesting way to reduce work hours, since they also eliminated one day of commuting. If everyone commuted 20 percent less, it would save a huge amount in commuting costs. And, not only would people be commuting fewer days, but by having fewer trips, we would have less congestion and less energy wasted by cars sitting in traffic. That sounds like a really good deal all around. (Of course, working from home, and not commuting at all, is even better.)
One issue is what happens to the productivity of workers who work a four-day week rather than a five-day week. The New York Times had a piece on an experiment in the United Kingdom, where 70 companies switched to a four-day week at the start of the year.
According to the piece, a private foundation covered the cost of paying workers for a fifth day, even when they were only coming in four days a week. The piece reports that the experience has been overwhelmingly positive, with the vast majority reporting that they intend to stick to a four-day workweek even after the experiment is over.
While this is very positive news for fans of a four-day workweek, the piece is very unclear about its measure of productivity. According to the piece, productivity did not fall at the companies that switched to a four-day workweek. But it is not clear how productivity is being measured.
Ordinarily, economists measure productivity as output per hour of work. Is the piece using this measure of productivity? This means that if workers were putting in four eight-hour days, then their output would be 20 percent less than when they were working five eight-hour days. If this is the case, it would be difficult to see how employers could pay them the same amount per week, unless they had extraordinary profits before the experiment.
It’s possible that workers are putting longer days now. Perhaps they are working 10-hour days, so their four-day workweeks still correspond to a 40-hour workweek. In that case, we can be encouraged that the longer day didn’t mean any decline in productivity, and if workers prefer one fewer day of work per week, that would be great.
It’s also possible that workers are putting in fewer hours in their four-day week than they did in their five-day week, but still managing to produce the same amount of output. There is some evidence than when France adopted a 35-hour standard workweek in the late 1990s it was associated with an increase in the rate of productivity growth. Perhaps we are seeing the same story with a four-day week.
However, to assess the extent to which productivity might have been increased in the companies switching to a four-day week, we have to know how many hours workers are putting in each day. It is not at all clear from this NYT piece (or the linked site for the organization coordinating the experiment), how many hours people are working.
If we take the extreme case, where workers are still putting in eight-hour days, the four-day week would imply a 25 percent increase in productivity. This is almost impossible to imagine. Economists would be thrilled by any policy that would increase aggregate productivity by just 1.0 percent over a period of years. That would imply an increase of $250 billion a year in annual output, that would be a really big deal.
A policy that could lead to an increase in productivity of 25 percent, overnight, is impossibly great. Even if workers were putting in nine-hour days and maintaining the same level of output it would imply an extraordinary 11.1 percent increase in productivity. That would be a really huge deal, as would an increase in productivity of even half this size.
Anyhow, it is hugely important for those advocating four-day workweeks to know what its impact is on productivity. Unfortunately, this piece provides no real basis for making that assessment.
I was happy to hear that my friend, Juliet Schor, is the lead researcher on this project. They are in fact making an effort to measure productivity, or at least revenue. Since many of these companies produce software, that should be pretty much the same thing.
She has informed me that while the bulk have gone to four eight-hour days a few have increased hours per day, so that they have not gone from a 40-hour week to a 35 or 36 hour-week. This would still imply very substantial productivity gains if they can maintain output levels.
According to Schor, even the ones that have gone from 40 to 32 hours are still managing to maintain the same level of output. They did this by planning for several months in advance and finding tasks (mostly meetings) that could be eliminated without reducing output.
In any case, Schor is a serious researcher and I’m confident that we will get useful data from this experiment after it is completed at the end of the year. Hopefully, the New York Times will do a follow up piece when the final report is available.