Many service sectors rely heavily on young and middle-aged workers, but these workers are increasingly scarce in aging economies. As many as 18 out of 36 countries in the Organisation for Economic Co-operation and Development will experience declining populations by 2055. Robots, originally developed for manufacturing firms, have become increasingly common in service organizations. However, little is known about how robot adoption in the service sector affects employment, tasks, and productivity. Though robots often prompt concerns about job replacement, countries with aging populations such as Japan increasingly view robots as key to augmenting a declining working-age population and alleviating the shortage of workers. Our research studies the effects of robots on labor and service quality in Japanese nursing homes.

Worker shortages have been especially salient in elderly care. In many countries, nursing homes experience persistent staff shortages and high levels of turnover while the elderly population and demand for caregiving grow. Caregivers assist elderly residents with daily activities—such as eating, exercising, bathing, and toileting—and provide specialized care to residents with dementia or limited mobility. Thus, caregiving in nursing homes is a physically and emotionally challenging job, but the pay is lower than that of many service-sector jobs that do not require extensive skills. The COVID-19 pandemic and the high number of cases and deaths in nursing homes have underscored the challenges of consistent, high-quality staffing and care for the elderly.

We conducted two surveys of Japanese nursing homes in early 2020 and 2022 to collect detailed information on robot adoption, staffing, quality of care, the adoption of other technologies, and management practices. In the second survey, we also collected information on the extent to which nursing homes used robots for various tasks and categorized robots based on their function.

Some models of automation and jobs suggest that automation technologies replace tasks performed by humans. However, few studies have collected detailed data to scrutinize this model. We identified the key tasks performed by caregivers in nursing homes and analyzed how the share performed by robots changed between 2020 and 2022. Finally, we collected data on the staffing levels, experience levels, and wages of two types of nursing home workers: regular workers, who have stable employment contracts with benefits, and nonregular workers, who often work part-time and have more flexible contracts with fewer benefits.

First, our research finds that robot adoption was positively associated with the number of caregivers and nurses at nursing homes, especially nonregular workers. Specifically, a 10 percent increase in robots used in nursing homes was associated with a 0.24 percent increase in total employment and an approximately 0.3 percent increase in caregivers. The effect was greatest when facilities adopted monitoring robots—robots that assist caregivers by providing real-time information about residents and that assist residents in communicating with caregivers and family members. The average number of monitoring robots was 7.7 per nursing home in 2020; doubling the number of monitoring robots was associated with a 3 percent increase in caregiver staffing, equivalent to about 1.2 additional caregivers or nurses. This effect could be driven by robots alleviating caregivers’ physical burdens and demands on their time. Indeed, the rate at which nursing home employees left their jobs decreased as facilities adopted more robots. This enabled facilities to retain more workers, especially nonregular workers.

Second, robot adoption was associated with an increase in the share of each task performed by robots. Specifically, we analyzed communication tasks, mobility tasks, and transfer tasks. Additionally, higher levels of task automation were associated with increased employment of caregivers, especially nonregular caregivers.

Third, monitoring robots and transfer robots—robots that assist caregivers in lifting, repositioning, and transferring residents—were positively associated with the number of residents receiving care, nursing home revenue, and the length of a facility’s waitlist. Furthermore, the use of restraints and cases of bed sores decreased with robot adoption, especially monitoring robots, indicating that the quality of care and residents’ well-being improved.

Adopting robots in nursing homes may have improved productivity because robots enable workers to devote more time and attention to caregiving tasks. Moreover, previous research has found that high caregiver turnover is associated with lower quality of care. Decreased turnover can enhance the continuity and timeliness of care by providing more knowledge to caregivers of residents’ needs and by reducing gaps in service. However, our research also finds that mobility robots—robots that assist residents with daily physical activities—were associated with more falls. This suggests that problems can arise when vulnerable residents attempt to use new technologies to become more mobile.

NOTE
This research brief is based on Yong Suk Lee, Toshiaki Iizuka, and Karen Eggleston, “Robots and Labor in Nursing Homes,” Labour Economics 92 (January 2025).