INTERACTIVE FIGURE

Employee Outflow from Microsoft

One effect of the COVID-19 pandemic is that many workers who formerly worked five days a week in an office now work from home all or some of the time. And this trend will likely continue. As of June 2023, an estimated 75% of tech and 28% of all companies in the United States were still “fully flexible”—either fully remote or with a voluntary in-office option.

That said, employers and employees are often at odds on this development, with some employers even reinstating full return-to-office (RTO) policies. What are the effects of these decisions on the labor market? If employees prefer hybrid or full work-from-home options, will they leave employers who institute RTO policies? If so, which type of employees are more likely to leave, and where will they land?

To address these and related questions, the authors study the effects of RTO mandates at three large tech companies—Microsoft, SpaceX, and Apple. Together, the three companies account for over 2% of employment in the tech sector and 30% of its revenue. In other words, what happens at these companies matters for the American economy and sets the precedent for the wider debate around the return to office. Because these companies initiated their RTO mandates before a wave of layoffs hit the tech industry from late 2022 onwards, the authors were able to disentangle the causal effects of the RTO mandates. The authors’ novel methodological approach, which allows them to study in detail what happens within each company, reveals the following:

• Following the return-to-office (RTO) mandate, the authors estimate a reduction in tenure among employees that escalates with tenure duration; in other words, long-tenured and more senior employees are more likely to leave.
• Importantly, these senior employees are leaving to join direct, large competitors, and not start-ups. This means that RTO mandates not only add significant human capital costs from the hiring and training costs required to replace senior employees, but they also add the competitive costs of lost operational knowledge.
• The authors do not find evidence of distributional differences in the share of men vs. women who leave for other companies; the share of leavers that flow into unemployment; the share of leavers that accept a demotion in their title at their new job; or the share of leavers changing roles at their new job. Regarding gender differences, this is a bit of surprise given that recent surveys have suggested that women have a stronger preference for working from home.
• Finally, the lack of counterfactual changes in unemployment status, demotions, and job roles suggests that leavers have good outside options, given that they appear to find employment in similar roles and levels of seniority.

Bottom line: Companies that mandate return to office could face significant human capital costs in terms of output, productivity, innovation, and competitiveness for the companies that implement them. Any broader impacts on labor markets and the employment landscape are questions for future research.