SMU SOE Seminar (Oct 20, 2017): When Peers Count: Evidence from Randomized Peer Assignments in the Workplace

This paper explores how the ability of coworkers affects individual productivity. I study a seafood-processing plant in Vietnam, where workers process fish individually and are compensated based on own output. Input (steamed fish) is allocated in groups (worktables), which on average consists of four workers. Managers keep track of each group's progress to ensure that input is allocated according to its work speed. For a period of five months, workers were randomly assigned to worktables on a daily basis. Using random variation in coworker composition, I find that an increase in the ability of coworkers at the same worktable leads to a decrease in worker productivity, measured as kilograms of fish processed per hour. This productivity decline is shown to be mitigated when a coworker is close to the focal worker either in physical distance or through a social tie.