Quitting and peer effects at work

While peer effects have been shown to affect worker's productivity when workers are paid a fixed wage, there is little evidence on their influence on quitting decisions. This paper presents results from an experiment in which participants receive a piece-rate wage to perform a real-effort task. After completing a compulsory work period, the participants have the option at any time to continue working or quit. To study peer effects, we randomly assign participants to work alone or have one other worker in the room with them. When a peer is present, we manipulate the environment by giving either vague or precise feedback on the co-worker's output, and also vary whether the two workers can communicate. We find that allowing individuals to work with a co-worker present does not increase worker's productivity. However, the presence of a peer in all working conditions causes workers to quit at more similar times. When, and only when, communication is allowed, workers are significantly more likely to (1) stay longer if their partner is still working, and (2) work longer the more productive they are. We conclude that when workers receive a piece-rate wage, critical peer effects occur only when workers can communicate with each other.