Does Working from Home Improve Workers’ Performance?

Working from home (WFH) is an increasingly popular phenomenon in the US, as the nature of the workplace has changed drastically in the past decades. While WFH can be a great way to provide workers with more flexibility and positivity, there are also concerns about “shirking from home.” A recent National Bureau of Economic Research working paper by Nicholas Bloom, James Liang, John Roberts, and Zhichun Jenny Ying, entitled “Does Working from Home Work? Evidence from a Chinese Experiment,” attempts to study the effect of working from home through data of call center employees’ productivity collected at CTrip, the largest Chinese travel agency.

CTrip decided to run a nine-month randomized control trial to test the impact of WFH and determine whether the practice should be adopted firm-wide. While WFH would help the firm reduce office rental costs and attrition rate, it might harm the firm in terms of productivity if workers could not keep their focus when working without supervision. Employees in the airfare and hotel departments were asked whether they would be interested in working from home four days a week, with the fifth day in the office as usual. Among those interested and eligible, those with even-numbered birthdays were randomly selected to work from home and while the rest stayed in the office and acted as the control group.

The study found that the performance of the home workers increased 13 percent over the course of the experiment. The improvement came mainly from a nine percent increase in the number of minutes worked, which was attributed to a reduction in breaks and sick-days. The remaining four percent came from home workers increasing the number of calls per minute worked. Many workers associated this gain with quieter working conditions.

While the gains in the quantity of work accomplished were significant, the quality of work was not compromised. In addition, the attrition rate from people leaving the company fell sharply by 50 percent compared to the control group. Home workers also reported higher work satisfaction and more positive attitudes. Another benefit was that there were no negative spillovers: those in the control group did not perform worse than other similar employees, despite their loss of treatment status. The one important downside to WFH was that, conditional on performance, it reduced promotion rate by 50 percent.

Because of the large estimated savings of $2,000 per employee per year, CTrip extended the program to all workers. As the workers in the treatment and control group could re-select their work environment, half of them changed their minds, indicating that employees learned a lot about their suitability for WFH. The increase in performance of home workers now rose to 22 percent. However, this result is largely driven by selection bias: those more suited for WFH chose to do so while those more suited for office environments chose not to join the program.

As this study is the first randomized experiment on WFH, it provides us with some of the first formal evidence about the impact of WFH. However, the study has two major limitations. First, as the experiment only concerns call center workers, employees with lower wages and quantifiable work outcomes, the study’s results could not be extended to other types of workers, such as those with higher wages and hard-to-measure work outcomes. Second, as the workers were first asked whether they would be interested in WFH program before being assigned to either the treatment or the control group, those participating in the program might be very different from the employees in the rest of the company. The authors found that they were more likely to be married, have children, and face long commutes to work. This would affect the external validity of the program. Nevertheless, the significant increase in workers’ performance as shown in this study indicates the great potential for WFH and will hopefully motivate more research in this area.

Feature Photo: cc/(Kevin Galens)