Can a Voice Channel Reduce Turnover? Evidence on Employee Voice and Exit from a Cluster-randomized Trial in U.S. Fulfillment Centers

Abstract

Building on the long tradition of research on employee voice and its potential impact on both employee and organizational outcomes, we investigate whether a new voice channel reduces turnover in e-commerce fulfillment centers. A cluster-randomized trial compared hourly workers in sites randomized to launch the new voice channel (Health and Well-Being Committees, or HaWCs) with those employed by the same firm in control sites. This participatory intervention involved a small group of frontline workers and supervisors who solicited concerns and ideas about safety, work processes, and other workplace stressors from the broader workforce and then developed and implemented improvement projects in response. HaWCs were an isolated change, rather than one component of a broader high-involvement work system, and they were implemented in non-union worksites, raising questions about their likely impact. Using administrative data on the population of hourly workers, an intent-to-treat (ITT) analysis finds individual workers’ monthly probability of exit fell by 1.3%-points in HaWC buildings in the year after randomization, representing a 20% decline relative to pre-intervention exit rates. Workers in buildings where HaWCs completed more projects were less likely to exit, showing that voice can yield visible improvements in the work environment that help retain workers. HaWCs also reduced exits over and above both pre-existing and new, alternative channels for soliciting employee input, suggesting that additional benefits flow from the HaWCs’ participatory nature. These findings indicate the feasibility of addressing turnover and improving the work environment through employee voice, even in tough conditions like fulfillment centers.