A randomized controlled trial published in the leading academic journal Nature finds two days of working from home improved job satisfaction and reduced turnover when compared to those working in offices five days a week.
The long-simmering debate over whether remote work hurts productivity or dampens worker performance has reached the peak of academic prestige: A study in the journal Nature.
Leaders eager to get workers back to the office often believe it leads to things like lower worker output and an erosion of workplace culture. But a new paper to be published Wednesday in the leading academic journal by Stanford University economist Nicholas Bloom and his co-authors suggests new evidence for advantages to hybrid schedules. A randomized controlled trial of employees in a tech firm based in China found that two days of work from home and three days in the office reduced quit rates, improved job satisfaction and had no impact on performance, when compared with employees who worked full-time in the office.
“Everyone’s heard of Nature,” says Bloom. If you’re talking about flexible work arrangements with a CEO who wants people back full time, “having a large, randomized controlled trial—they can still ignore it, but it gets a lot harder.”
The study randomly divided 1,612 workers at Trip.com, a large global travel firm, by whether they had even or odd birthdates. One group was able to work from home on Wednesdays and Fridays; the other worked in the office all five days. The researchers found that for those with hybrid schedules, attrition rates dropped by one-third over the six-month experiment period—and were even greater for non-managers, female employees and those with long commutes—while work satisfaction scores improved. The hybrid group also did not have significant differences on performance reviews or promotion rates, even for up to two years after the start of the experiment, and there were not significant differences in the lines of code submitted by the software engineers between the two groups.
That trial method—a real-life experiment on two employee groups randomly split by nothing other than birthdate—allowed the researchers to conclude it was the hybrid work schedule that caused the improvements in retention and job satisfaction, Bloom says, rather than simply being coincidental to other factors. “In many studies you don’t [have that],” Bloom says. “There are no differences between people born on even and odd birthdays.”
The study did not look specifically at workers who are full-time remote, or at arrangements where employees have personal choice over where and when they work, which remains a priority for many workers. That debate will likely continue. But the current study offers additional evidence to support the value of employees getting to work from home at least part of the time as opposed to full-time in the office.
In the paper, Bloom and his co-authors note that some may wonder if a possible explanation for the reduced turnover in the hybrid work group is that members of the full-time office group were frustrated they didn’t get the flexible schedule during the experiment. However, turnover rates for full-time office workers were slightly reduced from the six-month period before the experiment, the researchers write in the paper, suggesting that some may have “guessed (correctly) that the policy would be rolled out to all employees once the experiment ended.”
Another effect: The study found that managers at the company shifted their views of hybrid work over the experiment period. Before it started, managers at the company perceived that hybrid work would reduce productivity by 2.6%; by its close, they believed flexible arrangements could improve productivity by 1%. “There is value in experimentation,” the study authors write.
The paper follows months of debate over the impact of remote work on company culture, employee productivity and potential downsides to innovation or collaboration. Bloom, who has studied remote work for decades, prompted a stir last year when a working paper that reviewed existing studies on the topic pointed to research showing that fully remote workforces appeared to have slightly reduced productivity on average. But the studies also showed that when the work schedule was well-managed and hybrid at least some of the time, the effect was flat or slightly positive, Bloom told Forbes at the time.
Indeed, an early paper by Bloom looked at the same company, Trip.com (one of its co-founders, James Liang, is a co-author and former PhD student of Bloom’s). It found that among a group of remote call center agents who worked in person just one day a week, productivity increased 13% and turnover fell by half.
The newer study looked at workers in fields like marketing, software engineering, finance and accounting, helping to address concerns that past studies on lower-paid employees doing repetitive tasks with more objective measurements may not be generalizable to the workforce at large. “These are creatives, they’re in graduate or professional jobs, and they’re innovating, creating, training,” Bloom says of participants in the current study.
Trip.com, Bloom says, was seeking a way to reduce costs in doing the experiment, estimating that each employee who quit cost the company $20,000 in recruiting and training, the paper reports. As a result of the study, the company decided to extend the hybrid policy to all employees.
“There are good and bad things about working from home, but it turns out with hybrid they roughly net each other out,” says Bloom. “Employees were dramatically happier if they get to work from home two days a week, and as a result their quit rates fell by a third. ... The company looked at this and said, ‘what’s not to like?’ ”
